Full text of "Detailed Abundances for 28 Metal-poor Stars: Stellar Relics in the Milky Way"

See other formats

Draft version June 1 8, 2008 

Preprint typeset using I4TgX style emulateapj v. 10/09/06 



00 

O 
O 
(N 

00 



6 



> 
O 

m 
o 

oo 
O 



X 



DETAILED ABUNDANCES FOR 28 METAL-POOR STARS: STELLAR RELICS IN THE MILKY WAY'-^ 
David K. Lai^, Michael BoLTE^ Jennifer A. Johnson"*, Sara Lucatello^-'', Alexander Heger^-', and S. E. Woosley^ 

Draft version June 18, 2008 

ABSTRACT 

We present the results of an abundance analysis for a sample of stars with -4 <[Fe/H]< -2. The data 
were obtained with the HIRES spectrograph at Keck Observatory. The set includes 28 stars, with effective 
temperature ranging from 4800 to 6600 K. For 13 stars with [Fe/H]< -2.6, including nine with [Fe/H]< -3.0, 
and one with [Fe/H]= -4.0, these are the first reported detailed abundances. For the most metal-poor star 
in our sample, CS 30336-049, we measure an abundance pattern that is very similar to stars in the range 
[Fe/H]'--^ -3.5, including a normal Ch-N abundance. We also find that it has very low but measurable Sr and 
Ba, indicating some neutron-capture activity even at this low of a metallicity. We explore this issue further 
by examining other very neutron-capture-deficient stars, and find that at the lowest levels, [Ba/Sr] exhibits the 
ratio of the main r-process. We also report on a new r-process-enhanced star, CS 31078-018. This star has 
[Fe/H]= -2.85, [Eu/Fe]= 1.23, and [Ba/Eu]= -0.51. CS 31078-018 exhibits an "actinide boost", i.e. much 
higher [Th/Eu] than expected and at a similar level to CS 31082-001. Our spectra allow us to further constrain 
the abun dance scatter at low metallicities, which we then use to fit to the zero-metallicity Type II supernova 
yields of iHeger & Woosleyl (l2008h . We find that supernovae with progenitor masses between 10 and 20 Mq 
provide the best matches to our abundances. 

Subject headings: stars: abundances — stars: Population II — supernovae: general — nuclear reactions, 
nucleosynthesis, abundances 



1. INTRODUCTION 

In recent years the number of discovered extremely metal- 
poor (EMP, [Fe/H]< -3.0) star candidates has grown sub- 
stantia lly, thanks in large part to the survey of iBeers et alJ 
1992h. and the more recent Hamburg/ESO (HES) survey 



Christlieb et al. i 2000). The high-resolution follow-ups to 



these s urveys (e.g . McWilliam et al. 1995; Cohen et al. 2004, 
2008]: ICavrel et al.l l2004t lAoki et al.i l2005b IBarklem et all 
2005h have verified about 100 stars with [Fe/H]< -3.0. 

These EMP stars play an import ant role in understanding 
the very first generation of stars (iBeers & Christliebl l2005l 
and references and discussion therein). The lower the metal 
content of a star, the fewer instances of nucleosynthesis and 
recycling that preceded its formation. For the most metal- 
poor stars we may have the opportunity to measure the undi- 
luted imprint of Population III nucleosynthesis. The best ex- 
amples of this possibilit y are the two mo st metal-poor stars 
known, HE 1 327-2326 (IFrebel et al.l2005b and HE 0107-5240 

' The data presented herein were obtained at the W.M. Keck Observatory, 
which is operated as a scientific partnership among the California Institute of 
Technology, the University of California and the National Aeronautics and 
Space Administration. The Observatory was made possible by the generous 
financial support of the W.M. Keck Foundation. 

^ This publication makes use of data products from the Two Micron All 
Sky Survey, which is a joint project of the University of Massachusetts and 
the Infrared Processing and Analysis Center/California Institute of Technol- 
ogy, funded by the National Aeronautics and Space Administration and the 
National Science Foundation. 

' Department of Astronomy and Astrophysics, University of Cali- 
fornia, Santa Cruz, CA 95064; david@ucolick.org, bolte@ucolick.org, 
alex@ucolick.org, woosley@ucolick.org. 

Department of Astronomy. Ohio State University, 140 W. 18th Ave., 
Columbus, OH 43210; jaj@astronomy.ohio-state.edu. 

' Osservatorio Astronomico di Padova, Vicolo dell'Osservatorio 5, 35122 
Padua, Italy; sara.lucatello@oapd.inaf.it. 

* Excellence Cluster Universe, Technische Universitiit MUnchen, D-85748 
Garching, Germany. 

' Theoretical Astrophysics Group, T-6, MS B227, Los Alamos National 
Laboratory, Los Alamos, NM 87545. 



dChristlieb et al.ll2002l) . Both stars have [Fe/H]~ -5.3, and 
their abundance ratios can be fitted by zero-metallicity su- 
pernovae (SNe ) with a tuned mixing parameter in the pre- 
ejecta material dlwamoto et al.i r2005). An alternative scenario 
to explain the HE 107-5240 abundance ratios is proposed 
by lSuda et aP (l2004 . in which it is a Population III stai- that 
has accreted its heavy elements through binary and interstel- 
lar medium (ISM) accretion. In either case we are most likely 
seeing the imprint of the first stars, whether it is in SN ejecta 
or asymptotic giant branch (AGB) evolved material. How- 
ever, t here is a third possibility put forth bv [Venn & Lambert 
(l2008l) . that these stars are of a class of chemically pecuUar 
stars that have true [Fe/H] values greater than -4.0. 

As the sample of EMP stars has grown, a curious feature 
of the metal-poor end of the metallicity distribution function 
(MDF) of st a rs in t he Galaxy has become apparent. Recently, 
iNorris et all ( l2007h discovered a star with [Fe/H]= -4.8, HE 
0557-4840, making a total of only three stars with [Fe/H]< 
-4.1. Howe ver, when including these stars in the MDF, as 
iNorris et"aD (2007) point out, the number of stars discovered 
with metalHcities -5.3 <[Fe/H]< -4.1 still falls 3-4 times 
short of what is expected from the mixing and fallba ck models 
of chemical enrichment (e.g..' Salvadori e t al."2007'). Instead, 
a two-component model, as described by Karlsson (2006), 
where feedback effects from the first massive stars inhibit 
more star formation, may better fit the statistics of the ob- 
served halo MDF. 

One key to understanding these issues is to increase the 
sample of well-studied EMP stars to the point where the dif- 
ferent classes of abundance patterns can be identified and 
therefore explore the nature of their progenitors. While 
most EMP stars with [Fe/H]> -3.5 have relatively small 
dispersions for ele men ts at or below th e iron-peak (e.g. 
ICarrettaetaH 120021 and ICayrel et al.ll2004h . recent observa- 
tions have shown that there are some objects that are ei- 
ther strongly enh anced or deficient in certain a-elements 
(lAoki et al.ll2007l and references therein). At the more metal- 



2 



poor end. lCohen et all (l2007h report a star with [Fe/H]-- -4.0 
and a highly unusual abundance pattern, HE 1424-0241. It 
has very low [Ca/Fe] and [Si/Fe], -0.58 and -1.01, respec- 
tively, but has a [Mg/Fe ] of 0.44, which is typical of BMP 
stars. Cohe n et aP (12007) find that there are no core-collapse 
SN models that can fit this abundance pattern. These impor- 
tant results re-emphasize the need to find more stars in this 
metallicity regime. Only then can we begin to find out what, 
if anything, is typical, and it is clear with HE 1424-0241 that 
we must be very careful extending trends from [Fe/H]> -3.5 
to lower metallicities. 

EMP stars also exhibit a wide dispersi on in the ratio of 
the neutron -capture elements to ir on ie.g. iMcWilUaml [19981: 
iHonda et al . 2004; F ran9ois et af]|2007h . By examining both 
the most neutron-capture rich and neutron capture-poor EMP 
stars we can shed light on the different processes that give 
rise to this dispersion. Beginn i ng wi th the discovery of CS 
22892-052 dSneden et al.lll996l l2003l) . the imprint of a uni- 
versal r-process pattern has been found to stretch from EMP 
stars up to the Sun. As more r-process stars are discov- 
ered, this result has been even more strongly confirmed (at 
least for elements with Z> 56). For the i-process we are 
also beginning to see a convergence between observational 
abundance ratios of the neutron-capture e lements in these 
stars and models of EMP AGB stars (e.g. Johnson & Boltd 
|2b02a, 2004; Masseron et al. 2006). Even though some tun- 
ing of the '^C pocket formation in the AGB star is needed, the 
match between observed abundances and models is encourag- 
ing. These agreements, however, for both the r-process and 
i-process, do not explain the origin of some of the lighter 
neutron-capture elements (Z< 56). To investigate this, the 
measurement of the light neutron-capture element strontium 
may prove an ideal probe, as its resonance lines are still de- 
tectable in EMP stars. Part of the answer to this puzzle may 
come from looking at the most neutron-capture poor stars, to 
isolate the other process(es) that contribute to these elements. 

In this study we present abundance ratios from [C/Fe] to 
[Eu/Fe] for stars in various evolutionary states in the metallic- 
ity range -4 <[Fe/H]< -2. In addition to the various nucle- 
osynthesis events described above, mixing that occurs as a star 
evolves from the main sequence and up the giant bra nch can 
also £ t ffect the hght elem ent abundances up to N ( Gratto n et alj 
l2000HSpite et al.ll2005L 12006,) . Our data allows us to both see 
these evolutionary effects and provide a picture of the early 
Galaxy through its nucleosynthetic footprint. They also re- 
veal unexpected correlations of Teff with Si I, Ti I and Ti II, 
and Cr I. 

In § 2 and § 3 we present the details of the observations 
and analysis. In § 4 we present the abundance results of our 
study and compare them to previous samples of metal-poor 
stars. In § 5 we discuss interesting individual stars, as well 
as properties of the sample as a whole, including the curious 
behavior of Si I, Ti I, Ti II, and Cr I. Also in this section we 
present fits to the zero-m etallicity Type II SNe (Sne II) from 
iHeger & Wooslevl (l2008h . 

2. OBSERVATIONS AND REDUCTIONS 

We chose our sample from iLai et al.l (l2004h and metal- 
poor candidates ide ntified in the photometric sample of 
[Schuster et alj (|2004|) . The data w ere obtained from multi- 
ple runs at the HIRES spectrograph (IVogt et al.lll994 at Keck 
Observatory between 2001 and 2006. A detector upgrade in 
mid-2004 allowed us to obtain higher quality spectra in the 
blue region. Before the upgrade we typically observed using 



a blue and red configuration for HIRES, and after the upgrade 
we observed with a single setup. The details of the obser- 
vations, including wavelength coverage, signal-to-noise ratio 
(S/N), and V magnitude, are given in Table [T] 

The reductions were done differently for the data before and 
after the detector upgrade. The pre-upgrade spectra were re- 
duced with the MAKEE (Mauna Kea Echelle Extraction) re- 
duction package. The post-upgrade data were reduced with 
the HIRES data reduction package written by J. X. Prochaska. 

2.1. Equivalent Widths 

We used the spectrum analysis code SPECTRE 
dFitzpatrick & SnedenI 1198 7*) to measure individual equiva- 
lent widths (EWs) of isolated lines. The bulk of lines were 
measured by Gaussian fitting, but for some of the stronger 
lines we used a Voigt profile to better fit the line wings. 

As a check of our accuracy and data quality, we compare 
our EW measurements with previous high-resolution stud- 
ies done of commo n objects. F i gure [T] compares the EW s 
fro m the studie s of ICavrel et alj (l2004l) . lAoki et all (l2005b . 
and 'I vans et al ] (12001 " to our measurements. We share 63 
common lines with the Cayrel study for the star BS 16467- 
062, 438 common lines with the Aoki study from the four 
stars BS 16080-054, BS 16084-160, CS 30312-059, and CS 
30325-028, and 112 common lines with the Ivans study for 
the stars BDh-03 740 and BDh-24 1676. As can be seen 
from the figure, our values are in very good agreement with 
the three studies. On average we are finding slightly lower 
EWs than these previous studies of 1 .09, 2.41, a n d 2.24 mA, 
as compared to ICavrel et alj ( |2004|) . lAoki et alj (l2005h . and 
llvans etalJ(l2003ir respectively. 

3. STELLAR PARAMETERS AND ANALYSIS 

We used a cur rent version of Turbospectrum 
varez & PlezI Il998h . whic h properly accounts for 
continuum scattering (see lCavrel et a l. 2004) , in comb i nation 
with the stellar atmospheres from ICastelli & Kuruczl (l2003h 
to perform LTE line analysis and spectral synthesis. Our 
atomic line data along with measured EWs are giv en in Tables 
|2] through |4] We began with the line lists from llvans et aP 
(I2OO6) and iSneden et alJ (120031) . and added additional lines 
found using the NIST atomic line database. Specifically we 
updated/adde d the following g/-va lues: Fe I and Fe II lines 
with those of 'F uhr & Wiesd (I2006'). the Mg fine s at 5 172.7 
and 5 183.6 A with the values fromAldenius et al. ( 20071) . Cr I 
lines with v alues from Sobec k et al.l (l2007l) . Cr II lines with 
values from iNilssoiTet al.. (.20061), the Mn I Unes at 3577.9 
and 40 55.6 A with values from iBlackwell- W hitehead et aD 
(l2005h . and Zr II lines with values from Liung et al. (2006), 
when available, or otherwise from Malcheva et al. (2006i), 
when available. 

To measure oxygen, we adopted the linelist from Kurucz^ 
for the OH region at 3185 A with a dissociation potential of 
4.40 eV. We measured nitrogen from the NH feature at 3360 
A. Following the prescription from iJohnson et all (12007) for 
this list, we used the Kurucz ^/-values divided by 2 with a dis- 
sociation potential of 3.45 eV. The CH linelist at 4300 A was 
derived from the LIFBASE database (courtesy of B. Plez). 

3.1. Hyperfine splitting 
^ http://kuracz.harvai'd.edu 



3 



The effects of hyperfine splitting (hfs) can greatly affect the 
derived abundance from strong lines of certain elements. The 
effect is a desaturation of strong lines, and therefore a larger 
equivale nt width than would be fou nd given the absence of the 
splitting. Prochaska & McWilliam (2000) noted the effects of 
hfs for Sc and Mn. We have taken the hfs parameters from Ku- 
rucz to account for these two elements. Vanadium and cobalt 
are also known to be affected by hfs; however, all of our lines 
are below 20 mA, too weak for it to change their derived abun- 
dances. We have adopte d the hfs parameters and isotope ratios 
of Ba fr omlMcWilliamI (11998) along with updated ^/-values 
used by llvans et aH (l2006h. T he hfs parameters for Eu are 
taken from lSneden et al.l (12003 ). The near UV- lines that we 
use to measure Cu are also known to be greatly affected by 
hfs (B ihain et al. ' 2004). We account for this by using the hfs 
parameters from Kur ucz and assuming the solar isotope ratio 
of ^^Cu to ^^Cu from lAnders & Grevessd (Il989h . 

3.2. Radial Velocities 

Radial velocities were determined by cross-correlating our 
program star spectra with high-S/N template stars using FX- 
COR.^ These template star spectra were taken during the same 
observing run as the respective program stars, and their ra- 
dial velocities were measured with individual strong absorp- 
tion lines (typically 50 lines). These results are given in Table 
[U The typical internal error from this procedure is only ^0.2 
km s"'. However this does not take into account the overall 
zero-point error, particularly given the diversity of our instru- 
ment setups. A more realistic estimate to the error in absolute 
radial velocity can be had by looking at velocities derived for 
the same object observed during different nights with differ- 
ent instrument setups, but on the same run (i.e. 2004 April 
7-9). We find that there is an rms of ~1.0 km s~' in these 
measurements. In our sample BS 16084-160 is clearly iden- 
tified as a radial velocity variable. Also BS 16550-087 is a 
likely radial velocity variable. 

3.3. Model Atmosphere Parameters 

3.3.1. Teff 

We obtained our effective temperatures using the V - K 
color of each star. The K magnitudes were taken from the Two 
Micron All Sky Survey. We then transformed the color using 
the updated Alonso et al. ( 19 96|ll999h color-Teff calibration 
given by Ramirez & Melendez ^2005^ The reddenings were 
taken from the Schlegel et al. (1998) dust maps, except for 
values of E(B-V) greater than 0.10. The Schlegel map may 
overe stimate reddening for higher values (,Arce & Goodrnanl 
Il999l) : we adopted Equation 1 from lBonifacio et al.l ( 20001 ) to 
account for this. 

We find a trend in Fe I lines with excitation potential (x) in 
many of our stars. This trend is usually accounted for by ad- 
justing the Teff. In Table |5] we give the value for the slope of 
the trends, the corresponding correlation coefficient (r), and 
the number of Fe I lines used, for different cuts on the min- 
imum X considered. Almost all of our stars have a negative 
slope when considering all Fe I lines, which implies that our 
Teff is too high. Unlike Cohen et al. (2008), we find that us- 
ing the 0.2 eV cut still leaves a statistically significant cor- 
relation between individual Fe I lines and their correspond- 

' IRAF is distributed by the National Optical Astronomy Observatories, 
which are operated by the Association of Universities for Research in As- 
tronomy, Inc., under cooperative agreement with the National Science Foun- 
dation. 



ing x-values. ICavrel et al.l (|2004|) found that a cut at 1.2 eV 
seemed to eliminate the trends they found, and it does seem to 
markedly reduce the trends an d their statistic al significance in 
this study. However, unlike in lCavrel et al.l (^004), small but 
marginally significant trends still exist. In Figure |2] we show 
this effect for the above cuts on x- 

To test the accuracy of our Tgff determinations, we also fit 
the Balmer lines of two of our stars, CS 22880-086 and CS 
30336-049. CS 22880-086 in particular shows very Httle re- 
duction in the trends with the x cuts, and CS 30336-049 has 
the largest trend with x when considering all Fe I lines. In 
neither star do we observe Ha, and the H/3 lines are posi- 
tioned on the edges of the echelle orders. We therefore fit 
H7 and H(5. For both stars these two Balmer lines are fit- 
ted quite well with the V - K-denved Teff. Part of the an- 
swer to this discrepancy between spectroscopic temperature 
with color and Balmer line temperatures may come from the 
three-dimensional (3D) hydrodynamical effects in these stars. 
ICollet etaP (l2007h recently examined the potential impact of 
this in red giant stars. For metal-poor stars they find a strong 
effect that would explain the direction of our trend but overes- 
timates the magnitude. For this study we have chosen to stay 
with the V-A'-derived temperatures and note that the effects 
of inaccurate Teff are minimized by looking at abundance ra- 
tios versus absolute abundances. 

3.3.2. Surface gravity and Microturbulent velocity 

The Teff was then used to determine the surface gravity. 
We used the isochrones from Kim et al.. (2002). with a- 
enhancement set to 0.3 and age to 12 Gyr. We chose not to 
adjust the log g spectroscopically to get the abundance of iron 
from Fe II and Fe I to match. This way we avoid the poten- 
tial non-LTE (NLTE) effects on the Fe I abundance giving us 
an erroneous log g. Even without adjustment our Fe I abun- 
dances are in generally very good agreement with the Fe II 
abundances. 

The final microturbulent velocity was determined spectro- 
scopically by eliminating any trend with EW versus abun- 
dance for the Fe I lines. Because the Teff-color calibration 
is dependent on metallicity, and therefore the log g as well, 
we iterated the above method until we settled on a metallicity 
within 0. 10 dex of our Fe I abundance. The final atmospheric 
parameters are listed in Table |5] Figure [3] shows the span of 
evolutionary states of our sample. 

3.4. Error Analysis 

Excluding systematic NLTE and ID versus 3D atmosphere 
effects, the uncertainties in our measurements come from 
three sources. The first comes from the error in the EW mea- 
surement (or in the case of synthesis, the error in the fit). The 
second comes from errors in the atomic parameters. In most 
cases we have multiple lines measured for the same element 
in any given star and the scatter in those lines can give an esti- 
mate for the first two error sources. When four or fewer lines 
of an element in a star are measured, we calculate the aver- 
age dispersion for the sample in that element, and use this to 
set a minimum value. We then adopted the larger of the two 
values, the actual dispersion or this minimum value, to esti- 
mate this error component. If there is only a single line mea- 
sured for a given element, then we assume an error of 0. 15 in 
its abundance. In the case of the synthesis, we estimate this 
component of the error by how well fitted we can match the 
synthesis to the actual spectrum (this is typically 0.1). 



4 



The third source of error comes from the uncertainty in the 
model atmosphere parameters. We adopt the errors of lOOK 
and 0.2 dex for Teff and log g, respectively. The error estimate 
for Teff from using broadban d colors has been e stimated by 
multiple authors before (e.g. ICohen et al.l 12002). and lOOK 
is a conservative value. The log g error is more difficult to 
estimate. Although we do not use ionization balance to de- 
termine surface gravity, as mentioned above we still find that 
Fe I and Fe II agree very well in our stars. Changes to log 
g of 0.2 dex generally generate a noticeable difference be- 
tween Fe I and Fe II, and we use this as an error estimate of 
the isochrones themselves. We estimate the error for micro- 
turbulence velocity to be 0.2 km s"', as at differences greater 
than this pronounced trends of equivalent width versus Fe I 
line abundances appeared. 

We have adopt e d the error analysis techniq ue described by 
iMcWilUam et all (1 19951) and iJohnsonI (120021). In particular, 
we use equations 3, 5, and 6 from lJohnsonI ( |2002|) . including 
the covariance terms to take into account the dependent na- 
ture of our Teff, log g, and microturbulence values. Because 
our sample spans a wide range of evolutionary states, we 
use three different stars to estimate the atmospheric effects, 
BDh-03 740, CS 31078-018, and CS 29502-092, to cover the 
main-sequence/turnoff, sub- lower giant branch, and the up- 
per giant branch, respectively. In Tables |2l|8] and|9]we sum- 
marize the results for these three situations. Using equation 5 
and 6 from Johnson (2002), these values can then be used to 
estimate both the final absolute and relative errors. 

3.5. Comparison to Previous studies 

As a final check of our method and analysis we compare 
our atmosphere and abundance results to those stars from the 
studies listed in § 12.11 For B S 16467-062, we als o consider 
additional abundances from [B onifacio et al.l (12007.) and the 
abundance analysis bv lCohene t al. (20(3^! 

In Table |6] we summarize the atmospheric parameters. In 
general we are in very good agreement with these previous 
studies. The agreement with Cohen et al.l (I2008D is not sur- 
prising given that we derive our parameters in a very sim- 
ilar way. All of the other studies, however, use ionization 
balance to estimate surface gravities. That we agree with 
these log g values adds confidence to the isochrone method 
that we use to derive surface gravit y. One exception t o this 
good agreement is for BDh-03 740. Ilvans et al] (l2003b . pro- 
pose two atmospheres for this star. We have chosen to list 
in Tabl e [6l the atmosphere that most closely resembles ours. 
Ilvans e t al. (2003) go into detail about various other atmo- 
sphere and abundance determinations for this object. 

We show the abundance comparisons with these studies in 
Figure |4] In most cases we have good agreement, although 
there are a few exceptions. We find the largest discrepancy in 
[Al/Fe] compared to Aoki et al. (2005). The atomic parame- 
ters and at mospheres agree q uite well between this study and 
those from ! Aoki et al.l (120051) . so those are not the reason for 
the discrepancy. We believe the disagreement may arise from 
a CH absorption feature that is blended with the Al A3944 
line. 

Another highly discre pant ratio, this time between our study 
and that of Ilvans et al.l (12003 , ). is [Mn/Fe]. T his difference 
comes from measured EWs. Ilvans et aP (l2003l) measure only 
a single Mn I line at 4823.52 A to have EWs of 14.5 and 
7.3mA for BDh-24 1676 and BDh-03 740, respectively. For the 
same line we measure EWs of 3.0 and 1.2 mA. We have both 



higher S/N and higher resolution spectra than the Ilvans et al.l 
(l2003l) study, and for such low EW lines it is conceivable that 
their measurements for these two lines were overestimated be- 
cause of noise. This may also be the cause for the discrepancy 
we find in [Ni/Fe] . For both objects the co mmon lines are very 
weak, and we find much lower EWs than Ilvans et al.l (I2b03i) . 

4. RESULTS 

In the following section we discuss our abundance results 
along with those from various other high-resolution studies. 
These results are reported in Tables \\Q \ through [TSl as sum- 
ing the solar abundances from lGrevesse & Sauvall ( ll998l) . We 
comment on the specific elements below, but note that in gen- 
eral the abundances from all of these studies, including our 
own, agree remarkably well both in trends and in scatter. 

4. 1 . The light elements: C,N, O, and Li 

Figure |5] shows our values of loge(Li), [C/Fe], [N/Fe], and 
JO/Fe1 versus [Fe /H]. Overplotted on each are the values from 
ISpite et aLl(l2005l) . For all three elements we find a significant 
scatter through the whole range of [Fe/H]. In our most metal- 
poor star, CS 30336-049, we find sub-solar [C/Fe]. 

Although we do find similar values of [O/Fe] as ISpite et al.l 
(.2005) . there should be a word o f caution i n the man- 
ner in which we measure [O/Fe]. ISpite etal.1 (l2005h use 
the forbidden O I line at 6300 A, while we u se the UV 
OH re gion at 3185 A. According to Asplund & G arcia Perez! 
(1200 ll) . taking into account 3D effects may reduce oxy- 
gen abundances derived from the OH lines by as much 
as 0.6 dex. However, these models seem to over-predict 
the solar oxygen abundance derived from helioseismology 
(Delahaye & Pinsonneault 2006). Because of these uncertain- 
ties, we have chosen to present our results without any correc- 
tions. 

Because of limited wavelength coverage and unfortunate 
gaps from the pre-upgrade HIRES CCD, we are only able to 
measure the A6707 line of lithium for three stars. The lowest 
metallicity star with a Li abundance is the dwarf CS 22884- 
108, at [Fe/H]= -3.13, for which we find loge(Li)= 2.31. The 
other two measurements are for the stars CS 22872-102 and 
CS22878-027, with loge(Li) values of 1.97 and 2.39, respec- 
tively. The estimated error for all three measurements is 
0.17, and the average value for these three metal-poor dwarfs 
is 2.21. This agrees very well with the value reported by 
iBonifacio et al.l (l2007h . whose study of 17 metal -poor dwarfs 
find an average loge(Li)= 2.10. In Figure |5] it is clear that 
our values fall o n the Spite plateau, while the giants from 
ISpite et aH (l2005l) show a significant amount of depletion. 

4.2. Odd-Z elements: Na, Al, and Sc 

In 10 of our stars observed with the original HIRES CCD, 
our setup allowed us to measure Na from the Na D reso- 
nance lines at 5890 and 5895 A, and we were also able to 
measure Al in all but eight of our stars using one or both 
of its reson a nce li nes at 3944 and 3961 A. As noted by 
ICavrel et al.l (l2004l) . using these features introduces possible 
NLTE effects. We report our abundances without NLTE cor- 
rections. The suggest ed NLTE corrections wh en using these 
lin es are -0.5 for Na ("Baumu eller et al.lll998l) and +0.65 for 
Al (iBaumueller & Ge hren 199^^ 

As can be seen from Figure|6] we do not reprod uce the same 
trend in [Na/Fe] versus [Fe/H] as ICavrel et al] (|2004) . Al- 
though we have fewer measurements of [Na/Fe] in the higher 



5 



metallicity ran ge, our data suggest a flat trend in [Na/Fe]. 
Taken with the ICayrel et alJ (|2004) points, there appears to 
be a large scatter in [Na/Fe] down to [Fe/H]~ -3.25, and then 
little scatter for lower metallicities. 

Figure |6] also shows th at our [Al/ Fe] abundanc es agree re- 
markably well with the Cayrel et al.. (2004) and Cohen et al.l 
(1^04) distribution of values. In the metallicity range cov- 
ered by this study, we find a very low 0.12 dex dispersion in 
[Al/Fe]. 

4.3. Alpha elements 

The scatter in our a-elements, as shown in Fig. |7] is very 
small over the entire range of our metallicities. The rms scat- 
ter of these elements is very similar to the value found by pre- 
vious studies working in our metallicity regime. There are no 
truly anomalous outliers to the expected a-enhancement trend 
(e.g. [Aoki et al. 2007; Cohen et al. 2007). This suggests that 
all of our stars formed from gas produced with a very similar 
star formation history. 

For [Mg/Fe] and [Ca/Fe] the observed scatters are 0.13 and 
0.1 with average values of 0.32 and 0.31, respectively. We 
also find Ti II to be on average 0.09 higher than Ti I. However 
both the neutral and singly ionized species of Ti show rela- 
tively large scatter, about 0. 17 dex for both. We caution using 
this as a sign of true abundance scatter as we find a correlation 
of Ti with Teff. We discuss this further in § 15.21 

Another element that exhibits a large scatter is Si. As we 
show in Figure [8] we do not find any trends with metallicity 
and find a rms of 0.24 dex in [Si/Fe]. It appears that part of 
this scatter comes from a correlation of Si abundance with Teff 
whi ch w e also show in Figure[8] This is also discussed further 
in 



4.4. The Fe group (23<Z<28) 

Figure|9]shows the V abundances for our sample. Although 
we do not measure V I for many of our star s, we do not fin d 
an offset between V II and V I as found by JohnsonI (l2002l) . 
Overall both the neutral and ionized species give no trend with 
[Fe/H], reflecting the similar origi ns of V and Fe from exp lo- 
sive silicon and oxygen burning ( W ooslev & Weave3ll995h 

In Figure [TOl we plot both Cr I and Cr II as a function of 
[Fe/H]. We repro duce the dec l ining trend of [Cr I/Fe] with 
[Fe/H] as found in lCavrel et alj (|2004|) and references therein, 
albeit with a slightly steeper slope. Although we are only 
able to measure [Cr II/Fe] for our more metal-rich objects 
([Fe/H]> -3.3), we find a clear offset between the species, 
with an average [Cr Il/Cr I] of 0.22. Furthermore in the metal- 
licity range -3.3 <[Fe/H]< -2 that we measure Cr II, we find 
no evidence for declining [Cr II/Fe] with declining [Fe/H] . We 
explore this more in § 15.21 

Figure [TT] similarly summarizes our Mn abundances. The 
differences between our study and that of ICayrel et al.l (l2004l) 
arise from a correction that they chose to adopt for the abun- 
dances determined from the Mn I resonance triplet at 4030 
A. They note that in their objects these lines give a consis- 
tently lower value {^^ 0.4 dex) for Mn than the non-resonance 
lines. They therefore do not include them in the final Mn 
abundance. In the stars where only the triplet lines were de- 
tected, their abundance was adopted with a correction of H-0.4 
dex. We have chosen to include the Mn I triplet in our abun- 
dance determinations without correction, since it may have 
some unk nown dependence o n the atmospheric parameters of 
a star (the lCavrel et alj|2004l sample's concentration on giants 



avoids this problem). Whatever the cause of this discrepancy, 
it makes the Mn I abundances suspect. For this reason, we 
concentrate on the Mn II abundances. In Figure [TTI we have 
also plotted the Mn I abundances from Cayrel e t alj ( |2004|) 
on our Mn II abundances. As can be seen from the figure, 
our agreement is quite good, although we are finding slightly 
higher Mn abundances at [Fe/H]> -3.0. 

We generally confirm the increase i n [Co/F e ] with de- 
creasing [Fe / H] as found by McWilli am et alJ (Il995h and 
ICavrel et al.l (|2004|) . Ir i Figure [121 we se e that the trend 
matches the points from ICavrel et al.l (| 2004 |) in general scat- 
ter and slope. It is unclear why the points from lCohen et alJ 
(l2004l) are offset from ours. The line parameters for Co are 
similar among all three studies. Our [Ni/Fe] values are also 
shown in this figure, and they stay flat across all metallicities, 
with a very low rms of 0. 10 dex. 

4.5. CuandZn 

In Figure \T3\ we plot [Cu/Fe] versus [ Fe/H] and [Zn/Fe] 
versus [Fe/H]. It has been known since ISneden & Crocken 
( Il988l) that Cu is deficient in metal-poor stars. We are 
able to extend the [Cu/Fe] measureme nts down to JFe/H]= 
-4.0, comparable to the metallicities of ICohen et al.l (l2008l) . 
and down from the 



om tne p re' 
iMishenina etall l2002t 1 
2001 ! 



re vious low of 
immerer et al.l 



Fe/H] ~ -3.0 (e.g 



20031; iBihain et al.1 



T he abundance determ inations from ICohen et al.l (l2008l) 
and iBihain et al.l (|2004|) are also shown in the [Cu/Fe] 
pl ot. While we find a s imilar distribution o f [Cu/Fe] values 
as IBihain et all (|2004 . ICohen et all (l2008h measure higher 
[Cu/Fe] abundances. Part of this abundance spread is likely 
artificial. The version of MOOG used by Cohen et al. (2008), 
incorrectly treats continuum scattering as absorption. This 
is properly accoun ted for in the Turbospectrum code (e.g. 
ICavrel etal.ll2004 . We also derived [Cu/Fe] using MOOG 
and found that those values are on average 0.22 dex higher, 
with as much as a 0.6 dex difference. Our values are 
in good agreement with the chemical evolution models of 
iRomano & M atteucci (.2007) which assume an initial primary 
origin of Cu in SNell, and a secondary contribution from the 
weak s-process. 

We are only able to measure [Zn/Fe] in a handful of our 
stars due to gaps in the wavelength coverage of some of our 
spectra. The valu es that we are able t o measure agree nicely 
with the results of ICavrel et al.l (|2004|) . In particular, our data 
are consistent with the rise of [Zn/Fe] with decreasing [Fe/H] . 
Taken together, this rise in [Zn/Fe] could be indicative of an 
a-rich freeze out proc ess contributing at a higher level at low 
metallicities dCavrel et al. 2004). 



4.6. Neutron capture elements 

We are able to measure four neutron-capture elements in 
most of our stars: the light peak elements Sr, Y, and Zr (Z =38, 
39, and 40) and the heavy neutron capture element Ba (Z=56). 
We find a large scatter in these elements relative to Fe, as can 
be seen in Figure [T4l 

As found by other st udies (e. 
Johnson & Bolte 2002H iHonda et al 



McWi 


liam et al.l 


1995 


120041: 


Aoki et al.1 


2005 



iFranyois et al.i i2007i) . the light elements also appear to be 
highly correlated with each other, suggesting that these ele- 
ments share a common origin. In Figure [151 we show that 
[Y/Sr] and [Zr/Sr] have consistent values for all stars with 
their respective averages of -0.05 and 0.33. 



6 



The Ba resonance line abundances agree well with the non- 
resonance lines when measurable, and while we only have 
three stars with La measurements, these values are roughly 
consistent with the Ba measurements. The light neutron- 
capture elements, however, show a remarkable scatter rel- 
ative to Ba. Figure [15] also shows [Ba/Sr] versus [Fe/H]. 
The difference between the extreme valu es is almost 2 dex , 
replicating the finding of previous wor k dMcWilliamlll998l; 
iJohnson & Bolte 2002b; Honda et al. 200^7^ 

We also report on the discovery of a new r-process-rich star, 
CS 31078-018. Table [16] gives the summai-y of its Z> 38 
abundances, including thorium. In Figure [16] we show the 
spectral synthesis of the A4019 l i ne of Th II, using the log g f 
value measured bv lNilsson et alj (|2002|) . For most of the lines 
used to measure these neutron-capture elements hfs is negli- 
gible because of their low EWs (<30 mA). Except for Ba and 
Eu the only line that is noticeably affected in CS 31078-018 
is the resonance line used to measure Yb. To account for this 
we follow the prescription of Sneden et al. (2003) and use the 
hfs parameters from lMartensson-Pendrill et alj (I1994I) . 

5. DISCUSSION 
5.1. Mixing 

For some elements, we expect differences between giants 
and dwarfs because of internal mixing in giants. Models of 
mixing and observations of the effect on the abundance s in th e 
atmospheres of giants have a long history (e.g., Kraft I994|). 
Th e most relevan t obser vatio ns for very metal-p oor giants are 
the iGratton et all (l2000h and ISpite et al.1 (l2005h results. The 
mixing effects can be clearly seen in Figure[T7] where we have 
plotted the light elements (C, N, O, and Li) as a function of 
Teff, where we note that [C/Fe] drops off below ^5000K. We 
also plot [(CH-N)/Fe], which should remain largely unchanged 
with evolution, as a function of metallicity and Tgff in Figure 
[TSl There is no correlation of [(CH-N)/Fe] with [Fe/H] or with 
Teff, at least for the region for which we have more than upper 
Hmits, Teff below 5700K. 

In Figure [19] we plot loge(Li) as a function of luminos- 
ity for our sample a s well as those of iGratt on et al.' ('2000b 
for [F e/HK -1.3), ISpite et al.1 (l2005h . and Bonifacio et aQ 
2007h . We see the clear signature of the deepening convec- 
tive envelope. The average loge(Li) for the three dwarfs with 
Li measurements is 2.21 , consistent with the Spite plateau 
value of 2 . 1 as found by .Bonifacio et al.l (l2007h . 

5.2. Si, Ti and Cr, unexpected trends with Teff 

Other than for the elements discussed above, mixing pro- 
cesses and therefore the stage of stellar evolution that a star 
is in should not affect its abundances. However for the Si, Ti, 
and Cr abundances we do find a correlation of abundance with 

Teff- 

We plot our values of [Si/Fe] versus Teff in Figure [8] It 
is clear from this figure that there is a trend of decreasing 
Si with increasing Teff. The iPreston et al.l (12006 ) study of 
field horizontal branch stars and red horizontal branch stars 
in M15 also suggests this result, and shows that it is not cor- 
related with log g as well. This puzzling trend is unexpected 
from an evolutionary standpoint; the lack of correlation with 
log g sh ows that it is not tie d to mixing along the red giant 
branch. IPreston et al.l (l2006h carefully checked for contami- 
nation from possible CH lines and found very little effect, so 
that is unlikely the culprit. 

We also find a trend of Ti abundance versus Teff in the 



opposite sense of Si. As we show in Figure [20] this trend 
shows a decrea se in Ti with decre asing Teff- In Figure |20l 
the points from ICavrel et all (|2004 and ICohen et alT(l2004 
seem to confirm th i s tren d. We have also included the data 
from IPreston et a l. (2006) to fill in the sparse region in Teff 
between 5400 and 6000 K. With the i nclusi on of these hori- 
zontal branch stars from lPreston et al.l (l2006h . we see that like 
Si, this is predominately a Teff, not log g correlation. It also 
should be noted that this trend applies for both Ti I and Ti II, 
with respective slopes of 0.12 dex in [Ti I/Fe] per 500K and 
0.10 dex in [Ti II/Fe] per 500K. 

A similar trend to Ti is found in Cr I. In Figure |2T] we 
plot Cr I and Cr II as a function of Teff. A trend of declin- 
ing [Cr/Fe] with Teff can be seen, while [Cr II/Fe] appears, 
if not flat, then f ar less pronounced. Again we include data 
from Pr eston et al. (2006) in our plots, and we come to the 
same conclusion that this is a Teff and not a log g correlation. 
The slopes of the best fit lines for our data are 0.09 dex in 
[Cr/Fe] per 500K and only 0.03 dex in [Cr II/Fe] per 500K. 
This discrepancy in both slope and offset between Cr I and 
Cr II may point to NLTE effe cts. This has already been sug- 
gested bv lSobeck et al-l (l2007l) . and the case seems to be made 
even stronger by our data. Because Cr II appears relatively 
free of Teff trends, we performed a linear fit of [Cr Il/Cr I] 
with Teff and suggest that at least as a first step to correct Cr I 
abundances by this fit. We find that at 6500K, there is an off- 
set of 0.106 dex and a slope of -0.01 13 dex per 100 K. This 
leads to a correction of 0.3 dex at 4800K. 

A possible explanation for the observed trends with Teff 
is an incorrect T-t relationship in the adopted model atmo- 
spheres, in particular in giants. In fact, the adoption of a T-t 
relationship shallower than the true one would result in a de- 
rived abundance dependent on the depth of the line formation 
and hence on its strength, with strong lines yielding larger 
abundance values than weak lines. While this effect on the 
derived [Fe/H] abundances can be, at least partially, compen- 
sated by adjusting the value of the micro-turbulence, this does 
not apply to Si, Cr and Ti, as the abundances for the first come 
prevalently from lines forming the outer layers and that for the 
other two mostly from lines formed deep into the atmosphere. 
The observed trends could thus be explained in terms of in- 
creasing discrepancy between the model atmospheres and the 
"true" atmospheres at the decreasing of the stellar effective 
temperature. It is noteworthy that the derived slopes for the x- 
derived abundance relationship are steeper among giants than 
among dwarfs (see Table |5]l, which is what is expected in the 
hypothesis described. 

Whether these are ultimately due to true abundance trends, 
unaccounted for blends, or a deficiency in our knowledge in 
the spectral analysis (e.g. NLTE and other atmosphere ef- 
fects) i s yet to be determi ned. However by including the data 
from P reston et al.l (|2006), it seems that this is not an effect of 
stellar evolution. Regardless, these trends show that caution 
must be taken when looking at either Ti, Si I or Cr I to explore 
galactic chemical evolution or to constrain SN models. 

5-3- CS 30336-049, [Fe/H] =-A.Q 

Of our sample, the star CS 30336-049 is the most metal 
poor at [Fe/H]=-4.0. There are only three stars more metal- 
poor than this, and only about seven with comparable metal- 
licities (see Frebel et al. 2007c). It is comparable in metal- 
licity and atmosphe ric parameters to both CD-38:245 and 
CS 22949-037 (Cavrel etal.ll2004 . In its abundance ra- 
tios, CS 30336-049 looks far more like CD-38:245 flian CS 



7 



22949-037. Unlike for CS 22949-037, the derived [Mg/Fe] 
is actually slightly under-abundant at 0.08 compared to other 
metal-poor stars. While the high C,N,0 and Mg of CS 
22949-037 may be sug ge stive of a low-energy SN explosion 
(iTsujimoto & Shigeyama 2003), CS 30336-049 is "normal" 
in all of these abundances except for a high [N/Fe], leading 
to a low [C/N] of -1.2. At the temperature range of this star, 
~4800K, deep mixing may begin to occur and change some 
of the C to N. Although some of the N then may come from 
this internal processing in the star, it is possible that much of 
the low [C/N] value comes from the initial abundances of the 
cloud that formed the stars. The low carbon abundance of 
CS 30336-049, coupled with its very low [Ba/Fe] and [Sr/Fe] 
suggests that the star is not showing mass transferred material 
from an AGB companion, which suggests a primary source 
for its enhanced nitrogen abundance. 

Just comparing CS 30336-049 with CS 22949-049, it is 
clear that finding more stars at [Fe/H]= -4 is very important. 
Including HE 1424-0241 as discussed in the introduction, the 
picture gets even more complicated. If these are all prod- 
ucts of one to a small number of SN events, then a census 
of these objects may provide the best way to constrain the 
nature of Population III stars. With this in mind, we have at- 
te mpted to fit our abundan ces to the recent SN II model yields 
of lHeger & Wooslevl (1200 8) in § ET] 

The low-metallicity of CS 30336-049 also makes it a good 
candidate for constraining the gas cooling mechanism respon- 
sible for the Population III to Population II transition. As 
noted by lFrebel et al.l (l2007bh and references therein, there are 
two main competing ideas for this cooling mechanism follow- 
ing the initial metal enrichment from the first stars: atomic 
fine-structur e line cooling and d ust-induced fragmentation. 
Accordingly. lFrebel et alJ (l2007bh define a value they term as 
the "transition discriminant," £>txans' that is dependent on the 
overall C and O abundance of a star They find that the cooling 
from C II and O I fine-structure lines can allow for low-mass 
star formation only at values of Otrans > -3.5 ± 0.2. Intrigu- 
ingly, CS 30336-049 has a Dtrans value of -3.57. Given the 
inherent uncertainty of deriving O from the UV OH lines and 
the possibility that some C has been converted to N, this is not 
a highly certain value. However, if this value is correct, then 
this may be an indication that this star was formed due pri- 
marily to fine-structure line cooling from C and O produced 
from a Population III star. We note that this does not exclude 
the dust-induced fragmentation model, as even lower abun- 
dances of C and O can induce low-mass star formation in this 
scenario, but that now including CS 30336-049 all metal-poor 
stars with C and O measurements are consistent with the fine- 
structure cooling scenario. 

5.4. Neutron-capture-poor stars 

The existence of stars that are highly neutron-capture 
element deficient has been known for some tirne (e.g . 
iMcWilliam et"an[T995t iRvan et al J [19961: lMcWimam|[T99l . 
However these confirmed neutron-capture poor stars are rela- 
tively rare. In our own sample, BS 16084-160 and CS 30336- 
049 are exceptional in their low [Sr/H] and [Ba/H] values. 
Both of t hese values for BS 16084-160 agree with those re- 
ported by lAokietall (l2005h . iFulbright et"5] (|2004 analyzed 
the Draco dSph red giant D119 and found upper limits to 
[Sr/H ] and [Ba/HI that match our values. Th e recent stud- 
ies o fJCohe n etaL (2008). Fra ncois et aD(l2007h . lHonda et alJ 



(l2004jr and ,Aoki et al. (2005) have also added to a handful 
of stars in this regime. In Table [17] we summarize measure- 



ments from recent high-resolution studies for stars with both 
[Ba/Fe] < -1.0 and [Ba/H]< -4.0. 

In these neutron-capture element-deficient stars, a striking 
pattern emerges when comparing [Ba/H] to [Sr/H]. In Figure 
l22lwe can see that there seem to be two populations of stars. 
Below [Ba/H]^ -5.0, the Sr and Ba abundances appear to be 
well coupled. Above this value, however, there is a clear dis- 
tribution of production of Sr relative to Ba, with the trend of Sr 
being mainly over-produced relative to Ba. This suggests that 
the same process, such as a very low-level main r-process, is 
producing both the Sr and Ba in the hyper-neutron capture- 
poor stars (HNCP [Ba/H] < -5.0). 

To explain the productio n of Sr without much Ba, lead- 
ing to high [S r /Ba] r atios, lOian & Wasserburd (1200 lb and 
iTravaglio et alj |2004) propose a separation of the process 
that creates the light (e.g. Sr) and heavy (e.g. Ba) neutron- 
capture elements (termed the light ele mentary production pro- 
cess, LEPP, by Travagho et alj[2004 ). It has been proposed 
that the production site for these elements may come from 
charged-particle reactions in the neutrino driven wind off 
of a newly born neutron star (Oian & Wasserburg 20(37, and 
references therein). Recen tly both IMontes et alj (I20d7b and 
lOian & Wasserburg (l2007h explored how a two-component 
process could account for much of the scatter found in the 
light to heavy neutron-capture abundances. Figure l22l shows 
that at the HNCP end this LEPP may somehow be shut off, 
leaving only the signature of the main r-process. While there 
has been much concentration on neutron-capture rich stars, 
the observational constraints on the production site of the light 
neutron-capture elements will benefit greatly from more of 
these HNCP stars being discovered. 

5.5. CS 31078-018, A new r-process-rich star 

We have discovered a new r-p rocess-rich star, CS 31078 - 
018. Adopting the categories of 'Beers & ChristU eb" (20051), 
CS 31078-018 is an r-II star, that is, [Eu/Fe]> 1 and [Ba/Eu]< 
0. In Figure[23]we plot the neutron-capture abundances of this 
star over the solar sy stem r-process abundances taken from 
lArlandini et al] (Il999h and lSimmerer et all (|2004 . 

Similar to previously discovered r-process-rich stars, the 
agreement between the solar system r-process pattern and the 
abundances for CS 31078-018 is quite good for Z>56, further 
strengthening the case for a universal "main" r-process for the 
stable neutron-capture elements. 

We also find that CS 31078-018 exhibits an "actinide 
boost." The Th value is far higher than what should be ex- 
pected, given its radioactive decay lifetime and the theoreti- 
cal initial r-process production ratio. Our measured value of 
loge(Th/Eu)=-0.19 gives a negative age for the star if com- 
pared to the current estimate of -0.28 for the initial r-process 
production ratio (Kratz et al. 2007). This is the third star to 
exhibit this behavior , after CS 31082 -001 (iHiU et al.l[2002h 
and CS 30306-132 (Honda et alJl20CM l). 

In Fig. [24| we show the available measurements of 
loge(Th/Eu) from the literature for metal-poor stars. There 
are some clear uncertainties as shown by some of the dis- 
agreements between measurements for the same star, but it ap- 
pears that there is a real distribution of values of loge(Th/Eu), 
with CS 31078-018 near the top. Th e discrepancy between 
some of the results from lHonda et alj ([2004) and other mea- 
surements is at least partia lly explained by differences in their 
linelist with IWestin et alJ (^000), and differenc es in a d opted 
atmospheric parameters with Johnson & Bolt^ (1200 lb . lOianI 
([2001 proposes a mechanism that allows for a universal r- 



8 



process site for the heavy r-nuclei but can vary the actinide 
abundance via neutrino-induced fission, with neutrino expo- 
sure being dependent on individual SNe II. A distribution of 
values may lend credence to the idea of neutrino-induced fis- 
sion modifying the abundance of thorium. On the other hand, 
there seems to be an almost bimodal distribution of values, 
which is even more enhance d if we do not include the val- 
ues from lHonda et al.l (|2004|) . This is suggestive of two dis- 
tinct scenarios with values typified by CS 22892-052 and CS 
31078-018. Ultimately more [Th/Eu] measurements will be 
needed to settle this question. 

Regardless of the mechanism that over-produces the ac- 
tinides in some stars, it is clear from Fig. |24] that it does 
not affect all r-process-enhanced stars. Aside from this ac- 
tinide boost, there is no significant difference between these 
stars and normal r-process-enhanced stars, and chronometers 
based on comparing actinides to these stable elements must 
be approached with caution. 

5.6. Abundance Trends and Scatter in the Metal-poor 
Galaxy 

Fig. I25] summarizes the abundance measurements for our 
entire sample. In the metallicity range that we cover, the only 
elements to show a trend with [Fe/H] are Cr, Mn , Co, and Cu; 
our points are consistent with Cayrel et"aD (l200 4b for the trend 
they find in Zn. These abundance ratio trends are reflected in 
their larger scatters in Figure |25] As mentioned before, some 
of the sc atter for Cr I may be artificial since we also see a t rend 
with Teff, which also can explain why lCayrel et al.l (l2004l) find 
a much smaller scatter in Cr I as they have a much narrower 
Teff range. 

There is a true abundance spread in C and N, although as 
FigurefTSl shows, some of this can be reconciled by consider- 
ing [(CH-N)/Fe] (and therefore evolutionary state) instead. The 
large scatter in the neutron-capture elements, particularly be- 
low [Fe/H]=-3 as shown in Figure[T4]is also readily apparent 
in|25] 

What stands out from Figure |25] however is not just the 
scatter of the previously mentioned elements, but the remark- 
able consistency in other elements. For example the values 
for [Ca/Fe], [V/Fe], and [Ni/Fe], given reasonable errors, are 
effectively the same for all of our stars. This may also be true 
of [Si/Fe], [Ti/Fe], and [Cr/Fe], but this may be masked by 
their Teff dependence. We are seeing either the evidence of 
a very well mixed ISM when these stars formed, or evidence 
that the progenitor stars, possibly Population III stars, were 
all very similar. 

5.7. Abundance Pattern fits to zero metallicity SN 

If we assume that these early stars were formed from the 
products of the first stars, then we can try matching them 
to some of the most up-to-date nucleosynthesis results from 
zero-metallicity SN explosions. We use the recent models of 
[Heger & Wooslev (2008), which are updates to the models 
of iWoosley & Weaven (il995h . These models range in mass 
from 10 to 100 Mq, energies f rom 0.3 to 10.0 (x 10^' ergs), 
and mixing from none to 0.25. iHeger & WooslevI (1200 8) use 
a ID code and mixing cannot be followed due to its multi- 
dimensional nature. Instead, in these model s an artifical pre- 
scription for this parameter is used from Pinto & WooslevI 
([1988), and the amount of mixing is defined in terms of the 
ma ss fraction of the helium core (for a detailed explanation 
see iHeger & WoosieVi,2008) . We then assume that these stars 
are the product of one to just a few SN (e.g., iTumhnsonI 



I2006bh or assume an IMF of the first stars and match the yield 
from that to our stars. 

We first fit the average abundances of our sample to these 
models. There are a few general assumptions that we are us- 
ing to guide us. For all of our fits we are assuming that the 
Sc and Zn abundances derived from the models are lower 
limits. It is possible that part of the abundances of both of 
these elements are synthesized in pro ton-rich outflows from 
core-collapse SNe (iPruet et al.l l2005h . and p art of the zinc 
may also be made in a neutrino-powered wind dHoffman et alj 
1996). Neither of these pro cesses is included in the models 
of He ger & WooslevI dlOOSi) . We fit for Ch-N, instead of C 
and N separately to account for potential internal cycling, and 
we ignore O because of possible offsets due to using the OH 
feature. Finally we have added in the NLTE corrections dis- 
cussed in i^ l4.2l to our Na and Al abundances. 

We also find a curious result that the models always over- 
produce copper relative to the observations. In general we 
ignore copper in our fits, although we discuss it more below. 
We are also concerned with the trends of Ti, Si, and Cr I dis- 
cussed in § 15.21 as well as the offset between the neutral and 
ionized species of Mn. For Ti and Si, we weight their values 
by only half in the fits compared to the other elements. For 
Cr and Mn, we choose to use the averages from their ionized 
species to reduce the impact of potential NLTE effects on their 
respective neutral species. 

In the IMF models the explosion energy of the SN for a star 
of mass M is parameterized as 

As a convenient SN energy unit we use bethe, 1 B = 10''' ergs. 
Possible values of £0 in the model database are 0.3, 0.6, 
0.9, 1.2, 1.5, 1.8, 2.4, 3.0, 5.0, and 10.0 B, and can be 
-1,-0.5,0, and 1. 

In Figure l26l a) we show the best fit for a Salpeter IMF 
(r=1.35), for progenitor masses from 10 to 100 M0 (the en- 
tire mass range of the models). In the first case, we assume a 
standard IMF, where we set Zi'^'^P = and £0 = 1-2. The only 
free parameter is the mixing fraction. The fit is actually quite 
good, returning a standard mixing parameter of 0.16, com- 
pared to the canonic al value of 0. 1 used to explain the light 
curve of SN1987A ( Pinto & Wooslevl[T988l) . 

We relax the constraints on ii^P and £0 in Figure l26l b). 
We let the energy run over the ranges listed above, and only 
restrict £''^''p > 0.0, and we also let the progenitor lower and 
upper masses float. The match is noticeably better, with = 
0.69 for the best fit model. The £0 = 5. OB, with the explosion 
energy flat over all masses, £'^''p = 0.0, and no mixing. Also, 
even though we let the lower and upper bounds of the IMF 
vary, the best fit still used the entire mass range of models. 

Finally in l26l c). we let all the parameters range freely. This 
now includes letting F have the values -0.65, 0.35, 1.35, 2.35, 
and 3.35, to look at multiple Salpeter-like IMFs. We find a 
nearly identical fit in terms of with l26l b). The Salpeter 
power law exponent is still given by r=1.35, and the mixing 
parameter is a low 0.025. The characteristic energy is now 
much lower with 0.6B, and £'^''p = -0.50. The low explosion 
energy coupled with the negative value of £'^''p means that we 
are in effect reducing the importance of high-mass objects. 
The increased fallbac k in these objects when the explosion 
energy is lowered (see lZhang et al. 2008) causes very little of 
their metals to be ejected, in particular the innermost Fe-group 
elements fall back first. 



9 



We also constructed Gaussian IMFs to try to match our av- 
erage abundance distribution. The Gaussians are centered at 
11,12, 13.5, 15, 17, 20, 25, 35, 40, 50, 75, and 100 M©. The 
widths, in log(Mass), range from 0.025 up to 0.5 dex. The 
energies of the explosions are determined by Eqn. [T] with the 
same ranges of Eq and E'^^'^. The best-fit Gaussian IMF yield 
is shown in Fig. IZTl a). The fit both looks qualitatively similar 
to l26l c) and is also quantitatively very similar with a best-fit 
= 0.70. This Gaussian IMF is centered at 11.0 Mq, with 
a width of 0.3 dex (truncated at 10.0 Mq). The energies are 
defined by E'^"? = -1.0 and Eq = 0.6, and the mixing is 0.025 1 . 
The similarity in the fits is not surprising given that this trun- 
cated Gaussian IMF would look very similar to the Salpeter 
IMF of l26l c). We also fit our average abundance pattern to 
the entire library of single SN yields. This may be instructive 
to define a "typical" Population III star, even though clearly 
this is not the origin of the average abundances. We show the 
result in Fig. IZTl b). The fit is quite good, with a = 0.615. 
This best-fit star has a mass of 14.4 Mq, explosion energy of 

I. 8B, and a low mixing parameter of 0.015. 

We have also performed this best fit analysis on the most 
metal-poor star of our sample, CS 30336-049. In Figure |28] 
we show cases similar to those described above, assuming a 
Salpeter IMF (a), a Gaussian IMF (b), and a single star pro- 
genitor (c). We could not measure Cr II in this object, and 
instead we use the correction to Cr I proposed in § 15.21 and 
adjust Cr I by H-0.3 dex for the fit. Because of the inherent un- 
certainty of this abundance, we also weight Cr by a factor of 
1/2 as is done for Si and Ti. The other assumptions are as de- 
tailed in the beginning of this section. The Salpeter IMF has a 
difficult time fitting the abundance pattern of this star. The fit 
has a X- = 1 .914, a very high Eo = lOB, = -1 .0, T = 1 .35, 
mixing of 0.025, and ranges over all masses. The Gaussian 
IMF is best fitted by £0 = 1 •2B, £"=''p = -1 .0, a central mass of 

II. Mq, a width of 0.225 dex, and mixing of 0.025. The x^ 
is also a somewhat large 2.40. The most interesting case is the 
single star fit. The fit is excellent at x^ = 0.425, with M=10.9 
Mq, E=0.6B, and a very low mixing parameter of 0.01. This 
explosion in particular does an excellent job of reproducing 
the Fe-peak element pattern. 

As a final test, we also examined how well different param- 
eters fit our abundances, as opposed to looking only at the 
single best-fit model. All of the best fits presented above fa- 
vor progenitor stars with fairly low characteristic masses, ~ 
10-15 Mq. The characteristic explosion energies, however, 
are not as well constrained, with £0 ranging from 0.6 to 5. OB. 
The reality is that with so many models that can be compared 
to (16,800 total), it is relatively easy to find very good fits that 
are also somewhat degenerate in x^- With this in mind we 
show in Figure|29]the best 1000 single SN fits in terms of Eq 
and mass both for the average abundance ratio of our sam- 
ple and for CS 30336-049. The grid-like nature in these plots 
comes from the discrete values of the models. Overall the re- 
sults of the best fits from above are borne out. Although in 
both cases there are a small number of models that have very 
high explosion energy and mass, the vast majority of fits for 
our average abundance pattern have masses in between 10 and 
20 Mq and Eq less than 3. OB, while the results for the abun- 
dance pattern of CS 30336-049 show an even narrower range 
of masses of 10-15 Mq and typical Eq less than l.OB. The 
mixing parameter could not be constrained with this method 
because mass and energy have far greater impact to the fits, 
and there is not a favored mixing value when looking at the 
fits in aggregate. 



These results s eem to fit well with the findings of 
iTumlinsonI (l2006ah . which suggests that the characterstic 
masses of the first stars were between 8 and 42 Mq. This is 
argued from empirical constraints based on the non-detection 
of Population III stars in the Galactic halo, the Galactic halo 
MDF, and reionization , not from detaile d chemical abun- 
dances as done here. As lTumUnsonI (l2006ah notes, these num- 
bers are also close to the results of theoretical models of pri- 
mordial sta r formation that incorpo rate formation feedback 
effects (e.g. iBromm & Larsonll2004l and references therein). 

5.7.1. Copper, A cautionary note on the choice of explosion 
mechanism 

It is unclear why copper is so over-produced in these mod- 
els. Part of the solution may come from the choice of 
these models to locate the piston used to parameterize the 
explosion at an abrupt entropy jump where the entropy per 
baryon (S/NAk) is equal to 4, approximately at the base of 
the convective she l l of t he pre-SN object. Previously in 
IWooslev & WeaveJ (1 19951) . it was located where the electron 
mole number, Y^, decreased suddenly, which marked the edge 
of th e iron core. 

In i Heger & Woosleyl (l2008l) . there are also a small subset 
of models with explosion energies of 1.2 and 10 B calculated 
with the piston located at the boundary. To test the effect 
of the different piston locations, we show in Fig. [30f a) the 
best-fit single-star model to our average abundances using the 
Fe models (including Cu). This model has a mass of 25.5 Mq, 
£0 = 1-2B, and mixing of 0.1. In l30l b) we show their model 
with the exact same parameters, but with the piston located at 
the S/NAk boundary. It is clear that the location of the pis- 
ton can greatly affect the [Cu/Fe]. In the Kg model [Cu/Fe] is 
about -0.70, compared to the approximately solar value found 
in the same S/NAk model. This is not to say which explosion 
mechanism is correct, as any specification is a parameterized 
approach, but it does indicate that copper is a less than ideal 
element to use to constrain current SN models. 

6. SUMMARY 

We have presented the abundances from C to Eu of 28 
metal-poor stars covering a wide range of Teff. In the pro- 
cess we have found abundance trends with Teff of Si I, Ti I 
and Ti II, and Cr I that may be pointing to the deficiencies 
of our standard 1-D, LTE spectral analysis, or less likely, an 
unknown physical process for these elements. In either case 
the unexplained trends in Si I, Ti, and Cr I with Tgff show that 
we must be careful when using them to constrain models of 
galactic chemical evolution and models of SN yields. 

Our sample includes the discovery of a new [Fe/H]= -4.0 
star, CS 30336-049. In CS 30336-049, we find abundance 
ratios that track the trends from more metal-rich objects 
([Fe/H]~ -3.5), except for a mildly low [Mg/Fe]. These re- 
sults for CS 30336-049 show that some of the stars around 
[Fe/H]= -4.0 have a similar origin to these more metal-rich 
objects. However, the 10 well-studied stars of similar or lower 
metallicity show a diversity of abundances far greater than 
found in the more metal-rich stars. 

We have also discovered a new r-process enhanced star, 
CS 31078-018. Like other r-process-enhanced stars, it 
has a heavy neutron-capture (Z>56) abundance pattern that 
matches the scaled solar system r-process pattern. Interest- 
ingly, it has a much higher [Th/Eu] than most other r-process- 
rich stars, one that matches the value found in CS 31082-001. 
From figure |24l it is clear that there is a diversity of [Th/Eu] 



10 



values in r-process-rich stars. Whether it is a bimodal dis- 
tribution or continuous is not yet clear, and it will take more 
Th measurements to be sorted out. Regardless, it shows that 
chronometers based solely on this ratio need to be used with 
caution. 

We also explored the origin of the lighter neutron-capture 
elements (Z< 56) by examining stars that are highly deficient 
in these elements. By using stars in this sample and the litera- 
ture, we have found that HNCP stars ([Ba/H] <5) only exhibit 
the signature of the main r-process in their [Ba/Sr] abundance. 
This is in contrast to stars with slightly higher [Ba/H], which 
show a wide diversity of [Ba/Sr]. This result suggests that if 
there is a secondary process that produces these lighter ele- 
ments (i.e. Sr), then it does not operate in the most metal- 
poor regime. In determining the secondary physical process 
this may prove to be an important constraint if this continues 
to hold true as more HNCP stars are found. 

Overall, we find very little scatter in our relative abundances 
for elements in the Fe-group and lighter, and we have largely 
confirmed many of the trends of abundance with met allicity 
for sta rs with [Fe/H]< -2.0 that were detailed in Cavr el et alj 
(|2004|) . The low rms suggests either a well mixed ISM 
or a common origin for our stars. With this in mind, we 
have compared the average abundance pattern of our sam- 
ple with the zero-metallicity SN 11 nucleosynthesis models 
of Heger & Woosley ( 2008). These fits seem to indicate that 
metal-free SN 11 progenitors with masses ~ 10-20 M0 can 



match our abundances very well. This comparison was also 
done with the most metal-poor star of our sample, CS 30336- 
049, where we find that a slightly narrower range of progeni- 
tor masses ^ 10-15 Mq give the best matches to its abundance 
pattern. 



D.K.L. would like to acknowledge Chris Sneden, Ruth Pe- 
terson, Thomas Masseron, Bob Kraft, and Yong-Zhong Qian 
for useful discussions, insights, and advice. 

D.K.L., M.B., and J.A.J, performed this work with the sup- 
port of the National Science Foundation (AST-0098617 and 
AST-0607770). 

S.L. performed this work with the support of INAF cofin 
2006 and the DFG cluster of excellence "Origin and Structure 
of the Uniferse." S.L. would also like to thank R. Gratton for 
helpful discussion. 

A.H. performed this work under the auspices of the US 
Department of Energy at the University of California Los 
Alamos National Laboratory under contract W-7405-ENG-36 
and the DOE Program for Scientific Discovery through Ad- 
vanced Computing (SciDAC; DE-FC02-01ER41 176). S.W. 
received support from the National Science Foundation 
(AST-02611) and from the DOE SciDAC Program (FC02- 
06ER41438). 

Facilities: Keck:l (HIRES) 



REFERENCES 



Aldenius, M., Tanner, J. D., Johansson, S., Lundberg, H., & Ryan, S. G. 2007, 
A&A, 461, 767 

Alonso, A., Arribas, S., & Martinez-Roger, C. 1996, A&A, 313, 873 
Alonso, A., Arribas, S., & Martinez-Roger, C. 1999, A&AS, 140, 261 
Alvarez, R., & Plez, B. 1998, A&A, 330, 1109 

Anders, E., & Grevesse, N. 1989, Geochim. Cosmochim. Acta, 53, 197 
Aoki, W., Honda, S., Beers, T. C, Kajino, T., Ando, H., Noiris, J. E., Ryan, 

S. G., Izumiura, H., Sadakane, K., & Takada-Hidai, M. 2005, ApJ, 632, 

611 

Aoki, W., Honda, S., Beers, T. C., Takada-Hidai, M., Iwamoto, N., Tominaga, 
N., Umeda, H., Nomoto, K., Nonis, J. E., & Ryan, S. G. 2007, ApJ, 660, 
747 

Arce, H. G., & Goodman, A. A. 1999, ApJ, 512, L135 

Arlandini, C, Kappeler, P., Wisshak, K., Gallino, R., Lugaro, M., Busso, M., 

& Straniero, O. 1999, ApJ, 525, 886 
Asplund, M., & Garcia Perez, A. E. 2001, A&A, 372, 601 
Barklem, P S., Chrisdieb, N., Beers, T. C, Hill, V., Bessell, M. S., Holmberg, 

J., Marsteller, B., Rossi, S., Zickgraf, E-J., & Reimers, D. 2005, A&A, 

439, 129 

Baumueller, D., Butler, K., & Gehren, T. 1998, A&A, 338, 637 
Baumueller, D., & Gehren, T. 1997, A&A, 325, 1088 

Beers, T. C., Chiba, M., Yoshii, Y., Platais, I., Hanson, R. B., Euchs, B., & 

Rossi, S. 2000, AJ, 119, 2866 
Beers, T. C., & Chrisdieb, N. 2005, ARA&A, 43, 531 
Beers, T. C, Preston, G. W., & Shectman, S. A. 1992, AJ, 103, 1987 
Bihain, G., Israehan, G., Rebolo, R., Bonifacio, P, & Molaro, R 2004, A&A, 

423, 777 

Blackwell-Whitehead, R. J., Xu, H. L., Pickering, J. C, Nave, G., & 

Lundberg, H. 2005, MNRAS, 361, 1281 
Bonifacio, P., Molaro, P., Sivaiani, T, Cayrel, R., Spite, M., Spite, E., Plez, 

B., Andersen, J., Baibuy, B., Beers, T. C., Depagne, E., Hill, V., Eranfois, 

R, Nordstrom, B., & Primas, E 2007, A&A, 462, 851 
Bonifacio, R, Monai, S., & Beers, T. C. 2000, AJ, 120, 2065 
Bromm, V., & Larson, R. B. 2004, ARA&A, 42, 79 

Carretta, E., Gratton, R., Cohen, J. G., Beers, T. C, & Christlieb, N. 2002, 
AJ, 124, 481 

CastelU, E, & Kurucz, R. L. 2003, in lAU Symposium, Vol. 210, Modelling 
of Stellar Atmospheres, ed. N. Piskunov, W. W. Weiss, & D. F. Gray, 20P- 
+ 

Cayrel, R., Depagne, E., Spite, M., Hill, V., Spite, E, Fran9ois, P., Plez, B., 
Beers, T, Primas, E, Andersen, J., Barbuy, B., Bonifacio, P., Molaro, P., & 
Nordstrom, B. 2004, A&A, 416, 1 1 17 

Chrisdieb, N., Beers, T. C, Barklem, R S., Bessell, M., Hill, V., Holmberg, 
J., Korn, A. J., Marsteller, B., Mashonkina, L., Qian, Y.-Z., Rossi, S., 
Wasserburg, G. J., Zickgraf, E.-J., Kratz, K.-L., Nordstrom, B., Pfeiffer, 
B., Rhee, J., & Ryan, S. G. 2004, A&A, 428, 1027 



Christlieb, N., Bessell, M. S., Beers, T. C, Gustafsson, B., Korn, A., Barklem, 
R S., Karlsson, T, Mizuno-Wiedner, M., & Rossi, S. 2002, Nature, 419, 
904 

Christlieb, N., Reimers, D., Wisotzki, L., Reetz, J., Gehren, T, & Beers, 
T. C. 2000, in The First Stars: Proceedings of the MPA/ESO Workshop 
Held at Garching, Germany, 4-6 August 1999, ESQ ASTROPHYSICS 
SYMPOSIA. ISBN 3-540-67222-2. Edited by A. Weiss, T.G. Abel, and 
V. Hill. Springer- Verlag, 2000, p. 49, ed. A. Weiss, T. G. Abel, & V. Hill, 
49-+ 

Cohen, J. G., ChristUeb, N., Beers, T. C, Gratton, R., & Carretta, E. 2002, 
AJ, 124, 470 

Cohen, J. G., Chrisdieb, N., McWilliam, A., Shectman, S., Thompson, 1., 
Melendez, J., Wisotzki, L., & Reimers, D. 2008, ApJ, 672, 320 

Cohen, J. G., Chrisdieb, N., McWilliam, A., Shectman, S., Thompson, 1., 
Wasserburg, G. J., Ivans, I., Dehn, M., Karlsson, T, & Melendez, J. 2004, 
ApJ, 612, 1107 

Cohen, J. G., McWilUam, A., Christlieb, N., Shectman, S., Thompson, 1., 

Melendez, J., Wisotzki, L., & Reimers, D. 2007, ApJ, 659, L161 
Collet, R., Asplund, M., & Trampedach, R. 2007, A&A, 469, 687 
Cowan, J. J., Sneden, C, Buries, S., Ivans, 1. I., Beers, T. C, Truran, J. W., 
Lawler, J. E., Primas, E, Fuller, G. M., Pfeiffer, B., & Kratz, K.-L. 2002, 
ApJ, 572, 861 

Delahaye, F, & Pinsonneault, M. H. 2006, ApJ, 649, 529 

Fitzpatrick, M. J., & Sneden, C. 1987, in Bulletin of the American 

Astronomical Society, Vol. 19, Bulletin of the American Astronomical 

Society, 11 29-+ 

Fran9ois, P., Depagne, E., Hill, V., Spite, M., Spite, E, Plez, B., Beers, T. C, 

Andersen, J., James, G., Baibuy, B., Cayrel, R., Bonifacio, P., Molaro, P., 

Nordstrom, B., & Primas, E 2007, A&A, 476, 935 
Frebel, A., Aoki, W., Christlieb, N., Ando, H., Asplund, M., Barklem, R S., 

Beers, T. C, Eriksson, K., Fechner, C, Fujimoto, M. Y, Honda, S., Kajino, 

T, Minezaki, T, Nomoto, K., Norris, J. E., Ryan, S. G., Takada-Hidai, M., 

Tsangarides, S., & Yoshii, Y. 2005, Nature, 434, 871 
Frebel, A., Christlieb, N., Nonis, J. E., Thom, C, Beers, T. C, & Rhee, J. 

2007a, ApJ, 660, LI 17 
Frebel, A., Johnson, J. L., & Bromm, V. 2007b, MNRAS, 380, L40 
Frebel, A., Norris, J. E., Aoki, W., Honda, S., Bessell, M. S., Takada-Hidai, 

M., Beers, T. C, & Christlieb, N. 2007c, ApJ, 658, 534 
Fuhr, J. R., & Wiese, W. L. 2006, Journal of Physical and Chemical Reference 

Data, 35, 1669 

Fulbright, J. P, Rich, R. M., & Castro, S. 2004, ApJ, 612, 447 
Gratton, R. G., Sneden, C, CaiTetta, E., & BragagUa, A. 2000, A&A, 354, 
169 

Grevesse, N., & Sauval, A. J. 1998, Space Science Reviews, 85, 161 

Hauck, B., & Mermilliod, M. 1998, A&AS, 129, 431 

Heger, A., & Woosley, S. E. 2008, ApJ, submitted (astro-ph/0803.3161) 



11 



Hill, v., Plez, B., Cayrel, R., Beers, T. C, Nordstrom, B., Andersen, J., 
Spite, M., Spite, F., Barbuy, B., Bonifacio, P., Depagne, E., Franfois, P., 
& Primas, F 2002, A&A, 387, 560 

Hoffman, R. D., Woosley, S. E., Fuller, G. M., & Meyer, B. S. 1996, ApJ, 
460, 478 

H0g, E., Fabricius, C, Makarov, V. V., Urban, S., Corbin, T., Wycoff, G., 

Bastian, U., Schwekendiek, R, & Wicenec, A. 2000, A&A, 355, L27 
Honda, S., Aoki, W., Kajino, T., Ando, H., Beers, T. C., Izumiura, H., 

Sadakane, K., & Takada-Hidai, M. 2004, ApJ, 607, 474 
Ivans, 1. 1., Simmerer, J., Sneden, C, Lawler, J. E., Cowan, J. J., Gallino, R., 

& Bisterzo, S. 2006, ApJ, 645, 613 
Ivans, 1. 1., Sneden, C, James, C. R., Preston, G. W., Fulbright, J. R, Hoflich, 

R A., Carney, B. W., & Wheeler, J. C. 2003, ApJ, 592, 906 
Iwamoto, N., Umeda, H., Tominaga, N., Nomoto, K., & Maeda, K. 2005, 

Science, 309, 451 
Johnson, J. A. 2002, ApJS, 139, 219 
Johnson, J. A., & Bolte, M. 2001, ApJ, 554, 888 
— . 2002a, ApJ, 579, L87 
— . 2002b, ApJ, 579, 616 
— . 2004, ApJ, 605, 462 

Johnson, J. A., Herwig, F., Beers, T. C, & Christlieb, N. 2007, ApJ, 658, 
1203 

Karlsson, T. 2006, ApJ, 641, L41 

Kim, Y.-C, Demarque, R, Yi, S. K., & Alexander, D. R. 2002, ApJS, 143, 
499 

Kraft, R. R 1994, RASP, 106, 553 

Kratz, K.-L., Farouqi, K., Pfeiffer, B., Truran, J. W., Sneden, C, & Cowan, 

J. J. 2007, ApJ, 662, 39 
Lai, D. K., Bolte, M., Johnson, J. A., & Lucatello, S. 2004, AJ, 128, 2402 
Ljung, G., Nilsson, H., Asplund, M., & Johansson, S. 2006, A&A, 456, 1 181 
Martensson-Pendrill, A.-M., Gough, D. S., & Hannaford, P. 1994, 

Rhys. Rev. A, 49, 3351 
Malcheva, G., Blagoev, K., Mayo, R., Ortiz, M., Xu, H. L., Svanberg, S., 

Quinet, R, & Biemont, E. 2006, MNRAS, 367, 754 
Masseron, T., van Eck, S., Famaey, B., Goriely, S., Plez, B., Siess, L., Beers, 

T. C, Primas, F., & Jorissen, A. 2006, A&A, 455, 1059 
McWilliam, A. 1998, AJ, 115, 1640 

McWiUiam, A., Preston, G. W., Sneden, C, & Searle, L. 1995, AJ, 109, 2757 
Mishenina, T. V., Kovtyukh, V. V., Soubiran, C, Travaglio, C, & Busso, M. 

2002, A&A, 396, 189 
Montes, R, Beers, T. C, Cowan, J., Elliot, T., Farouqi, K., Gallino, R., Heil, 

M., Kratz, K. ., Pfeiffer, B., Pignatari, M., & Schatz, H. 2007, ArXiv e- 

prints, 709 

Nilsson, H., Ljung, G., Lundberg. H., & Nielsen, K. E. 2006, A&A, 445, 
1165 

Nilsson, H., Zhang, Z. G., Lundberg, H., Johansson, S., & Nordstrom, B. 

2002, A&A, 382, 368 
Norris, J. E., Christlieb, N., Kom, A. J., Eriksson, K., BesseU, M. S., Beers, 

T. C, Wisotzki, L., & Reimers, D. 2007, ApJ, 670, 774 
Norris, J. E., Ryan, S. G., & Beers, T. C. 1999, ApJS, 123, 639 
Pinto, R A., & Woosley, S. E. 1988, Nature, 333, 534 
Preston, G. W., Sneden, C, Thompson, I. B., Shectman, S. A., & Burley, 

G. S. 2006, AJ, 132, 85 



Prochaska, J. X., & McWilliam, A. 2000, ApJ, 537, L57 

Pruet, J., Woosley, S. E., Buras, R., Janka, H.-T., & Hoffman, R. D. 2005, 

ApJ, 623, 325 
Qian, Y.-Z. 2002, ApJ, 569, LI 03 
Qian, Y.-Z., & Wasserburg, G. J. 2001, ApJ, 559, 925 
— . 2007, Rhys. Rep., 442, 237 
Ramirez, I., & Mel^ndez, J. 2005, ApJ, 626, 465 
Romano, D., & Matteucci, F. 2007, MNRAS, 378, L59 
Ryan, S. G., Norris, J. E., & Beers, T. C. 1996, ApJ, 471, 254 
Salvadori, S., Schneider, R., & Ferrara, A. 2007, MNRAS, 381, 647 
Schlegel, D. J., Finkbeiner, D. P., & Davis, M. 1998, ApJ, 500, 525 
Schuster, W. J., Beers, T. C, Michel, R., Nissen, R E., & Garcfa, G. 2004, 

A&A, 422, 527 

Simmerer, J., Sneden, C, Cowan, J. J., Collier, J., Wooff, V. M., & Lawler, 

J. E. 2004, ApL617, 1091 
Simmerer, J., Sneden, C, Ivans, 1. 1., Kraft, R. P., Shetrone, M. D., & Smith, 

V. V. 2003, AJ, 125, 2018 
Sneden, C, Cowan, J. J., Lawler, J. E., Ivans, I. I., Buries, S., Beers, T. C, 

Primas, F, Hill, V., Truran, J. W., Fuller, G. M., Pfeiffer, B., & Kratz, K.-L. 

2003, ApJ, 591,936 

Sneden, C, & Crocker, D. A. 1988, ApJ, 335, 406 

Sneden, C, McWilliam, A., Preston, G. W., Cowan, J. J., Burris, D. L., & 

Armosky, B. J. 1996, ApJ, 467, 819 
Sobeck, J. S., Lawler, J. E., & Sneden, C. 2007, ApJ, 667, 1267 
Spite, M., Cayrel, R., Hill, V., Spite, R, Frangois, P., Plez, B., Bonifacio, P., 

Molaro, P., Depagne, E., Andersen, J., Barbuy, B., Beers, T. C, Nordstrom, 

B., & Prunas, F. 2006, A&A, 455, 291 
Spite, M., Cayrel, R., Plez, B., Hill, V., Spite, R, Depagne, E., Franjois, R, 

Bonifacio, P., Barbuy, B., Beers, T., Andersen, J., Molaro, P., Nordstrom, 

B., & Primas, R 2005, A&A, 430, 655 
Suda, T, Aikawa, M., Machida, M. N., Fujimoto, M. Y., & Iben, I. J. 2004, 

ApJ, 611,476 

Travaglio, C, Gallino, R., Amone, E., Cowan, J., Jordan, F., & Sneden, C. 

2004, ApJ, 601, 864 

Tsujimoto, T., & Shigeyama, T. 2003, ApJ, 584, L87 
Tumlinson, J. 2006a, ApJ, 641, 1 

— . 2006b, New Astronomy Review, 50, 101 
Venn, K. A., & Lambert, D. L. 2008, ApJ, 677, 572 

Vogt, S. S., Allen, S. L., Bigelow, B. C, Bresee, L., Brown. B.. Cantrall. 
T., Conrad, A., Couture, M., Delaney, C, Epps, H. W.. Hilyard, D., 
Hilyard, D. F. Horn. E., Jem, N., Kanto, D., Keane, M. J., Kibrick, 
R. 1., Lewis, J. W.. Osborne, J., Pardeilhan, G. H., Pfister, T, Ricketts, 
T., Robinson, L. B., Stover, R. J., Tucker, D., Ward, J., & Wei, M. Z. 1994, 
in Presented at the Society of Photo-Optical Instrumentation Engineers 
(SPIE) Conference, Vol. 2198, Proc. SPIE Instrumentation in Astronomy 
VIII, David L. Crawford; Eric R. Craine; Eds., Volume 2198, p. 362, ed. 
D. L. Crawford & E. R. Craine, 362-+ 

Westin, J., Sneden, C, Gustafsson, B., & Cowan, J. J. 2000, ApJ, 530, 783 

Woosley, S. E., & Weaver, T. A. 1995, ApJS, 101, 181 

Zhang, W., Woosley, S. E., & Heger, A. 2008, ApJ, 679, 639 



TABLE 1 
Observation details 



Star 




VRef. 






Observation 


Total Exposure 


Wavelength 


S/N 


S/N 


S/N 


ID 






(km s"' 


)* 


Date (UT) 


Time (s) 


Coverage (A) 


5560 A 


4500 A 


3435 A 


BD+03 740 


9.80 


1 


173.1 


(0.2) 


2004, Oct 7-8 


3900 


3020-5870 


425 


340 


278 


BD+23 3130 


8.94 


2 


-285.1 


(0.2) 


2006, Aug 19 


960 


3055-5893 


424 


245 


161 


BD+24 1676 


10.79 


1 


-237.8 


(0.2) 


2004, Oct 7 


3600 


3020-5870 


320 


288 


193 


BS 16077-007 


12.41 


3 


-38.3 


(0.2) 


2005, May 29-30 


2400 


3020-5870 


142 


111 


61 


BS 16080-054 


12.776 


4 


-121.2 


(0.2) 


2004, Apr 7 


1200 


3220-4655 




78 


12 








-119.3 


(0.3) 


2004, Apr 9 


1200 


5385-7665 


130 






BS 16080-093 


13.61 


5 


-205.2 


(0.2) 


2001, Aug 25 


5400 


3770-5280 




85 




BS 16084-160 


13.14 


3 


-123.2 


(0.2) 


2001, Aug 25 


7200 


3770-5280 




103 










-141.5 


(0.2) 


2004, Apr 9 


1800 


5385-7665 


140 












-126.6 


(0.2) 


2005, May 29 


5400 


3020-5870 




111 


34 


BS 16467-062 


14.09 


3 


-92.4 


(0.2) 


2003, Jun 6 


7200 


3770-5280 




106 




BS 16545-089 


14.44 


6 


-162.3 


(0.2) 


2004, Apr 8 


9000 


3220-4655 




114 


27 








-162.5 


(1.0) 


2004, Apr 9 


3600 


5385-7665 


100 






BS 16550-087 


13.76 


7 


-144.4 


(0.2) 


2003, Jun 6 


7200 


3770-5280 




65 










-148.3 


(0.2) 


2004, Apr 7 


1800 


3220-4655 




63 


8 








-148.9 


(0.6) 


2004, Apr 9 


1800 


5385-7665 


134 






BS 16928-053 


13.47 


8 


-81.5 


(0.2) 


2004, Apr 7 


7200 


3220-4655 




101 


11 








-80.9 


(0.3) 


2004, Apr 9 


2700 


5385-7665 


121 






BS 16929-005 


13.61 


8 


-51.7 


(0.2) 


2003, Jun 6 


1800 


3770-5280 




108 










-51.3 


(0.2) 


2004, Apr 8 


7200 


3220-4655 




137 


27 








-52.5 


(0.5) 


2004, Apr 9 


2700 


5385-7665 


117 






CS 22872-102 


13.65 


3 


-59.5 


(0.2) 


2004, Apr 7 


7200 


3220-4655 




138 


28 








-56.7 


(0.2) 


2004, Apr 9 


3600 


5385-7665 


134 






CS 22878-027 


14.41 


3 


-91.5 


(0.2) 


2004, Apr 8 


3600 


3220-4655 




60 


13 








-90.1 


(0.4) 


2004, Apr 9 


2700 


5385-7665 


87 






CS 22880-086 


14.41 


9 


-114.0 


(0.2) 


2003, Jun 6 


7200 


3770-5280 




71 




CS 22884-108 


14.24 


3 


-18.1 


(0.2) 


2004, Apr 8 


7200 


3220-4655 




93 


18 








-18.1 


(0.4) 


2004, Apr 9 


3600 


5385-7665 


101 






CS 22944-032 


13.28 


9 


17.7 


(0.2) 


2004, Oct 7-8 


7200 


3020-5870 


212 


147 


74 


CS 22957-022 


13.34 


3 


-33.8 


(0.2) 


2004, Oct 8 


3600 


3020-5870 


123 


95 


45 


CS 22963-004 


14.98 


3 


292.4 


(0.2) 


2004, Oct 8 


9000 


3020-5870 


93 


71 


29 


CS 22965-054 


15.10 


3 


-283.0 


(0.2) 


2001, Aug 25 


7200 


3770-5280 




61 




CS 29502-092 


11.895 


3 


-68.2 


(0.2) 


2001, Aug 25 


5400 


3220-4655 




121 


45 








-68.5 


(0.2) 


2001, Aug 25 


5400 


3770-5280 




115 










-69.3 


(0.2) 


2004, Oct 7 


4800 


3020-5870 


253 


141 


95 


CS 29506-007 


14.18 


3 


55.1 


(0.2) 


2004, Oct 8 


5400 


3020-5870 


85 


81 


34 


CS 29522-046 


12.74 


3 


-106.8 


(0.2) 


2004, Oct 7-8 


5400 


3020-5870 


185 


166 


91 


CS 30312-059 


13.101 


4 


-156.4 


(0.2) 


2004, Apr 8 


3600 


3220-4655 




134 


25 








-155.2 


(0.2) 


2004, Apr 9 


4800 


5385-7665 


169 






CS 30325-028 


12.889 


3 


-148.8 


(0.2) 


2005, May 29 


5400 


3020-5870 


146 


129 


51 


CS 30336-049 


14.048 


3 


-237.5 


(0.2) 


2004, Oct 7-8 


7800 


3020-5870 


94 


85 


31 


CS 31078-018 


13.211 


7 


81.3 


(0.2) 


2004, Oct 7 


7200 


3020-5870 


167 


90 


64 


CS 31085-024 


14.010 


3 


-322.0 


(0.2) 


2004, Oct 7-8 


9000 


3020-5870 


150 


100 


61 



REFERENCES. — (I),Hauck & Mei-milliod ( 1998); (2) H0,i; et al. (2000); (3) Sch uster et alj 120041) : (4) IAoki et all 120051) : (5) lBonifacio et alj I200CI1 : (6) 
ICohenetalJ 120041) : (7) lNoi-ris etal.a999.) : (8) iHonda et alj (20041) : (9) lBeers et ajj )200CI) 



Heliocentric velocity and estimated internal error. 



TABLE 2 

Equivalent Widths and Atomic Parameters (full version available in online edition) 



Wavelength 


Element 


logg/ 


EP 


BD+03 740 BD+23 3130 


BD+24 1670 BS 16077-007 BS 16080-054 


BS 16080-093 


BS 16084-160 


BS 16545-089 


BS 16550-087 


BS 16928-053 


5889.95 


11.0 


0.11 


0.00 


128.50 


173.50 




124.40 


25.00 


121.00 


145.70 


5895.92 


11.0 


-0.19 


0.00 


107.50 


141.00 




97.70 




94.50 


123.60 


3829.36 


12.0 


-0.21 


2.71 


160.70 


152.40 


140.40 


136.40 


60.40 


139.10 


144.20 


3832.31 


12.0 


0.15 


2.71 




187.30 


167.20 


157.70 




167.10 


180.20 


4057.52 


12.0 


-0.90 


4.34 


6.70 28.60 


14.30 


15.50 




2.00 


13.20 





TABLE 3 

Equivalent Widths and Atomic Parameters (full version available in online edition) 



Wavelength 


Element 


logg/ 


EP 


BS 16929-005 


BS 16467-062 


CS 22872-102 


CS 22878-027 


CS 22880-086 CS 22884-108 


CS 22944-032 


CS 22957-022 


CS 22963-004 


5889.95 


11.0 


0.11 


0.00 


89.60 




79.50 


83.20 










5895.92 


11.0 


-0.19 


0.00 


73.30 




63.90 


50.00 










3829.36 


12.0 


-0.21 


2.71 


118.30 


80.00 






79.50 


137.90 


133.10 


96.90 


3832.31 


12.0 


0.15 


2.71 


















4057.52 


12.0 


-0.90 


4.34 


11.60 




9.60 






17.10 


19.90 





TABLE 4 

Equivalent Widths and Atomic Parameters (full version available in online version) 



Wavelength 


Element 


logg/ 


EP 


CS 22965-054 


CS 29502-092 CS 29506-007 


CS 29522-046 


CS 30312-059 


CS 30325-028 


CS 30336-049 


CS 31078-018 


CS 31085-024 


5889.95 


11.0 


0.11 


0.00 








126.90 










5895.92 


11.0 


-0.19 


0.00 








104.90 










3829.36 


12.0 


-0.21 


2.71 


95.40 


136.70 




135.90 


157.20 


83.00 


152.00 


138.00 


3832.31 


12.0 


0.15 


2.71 












103.70 






4057.52 


12.0 


-0.90 


4.34 




18.80 


46.70 


16.40 


27.70 




20.90 


11.10 



TABLE 5 
Stellar Parameters 



AUHnes x > 0-2eV x> l-2eV 



Star 


[Fe/H] 


Teff 




yt 




X trend 


r 


Numb. 


X trend 


r 


Numb. 


X trend 


r 


Numb. 


ID 




(K) 


(cgs) 


(km s" 




(dex/eV) 




lines 


(dex/eV) 




lines 


(dex/eV) 




lines 


dU+\Jj /4U 


—2.75 


6557 


A oo 
4.Zo 




1 o 
l.o 


—0.04 


—0.49 


102 


—0.03 


—0.38 


91 


—0.03 


—0.27 


7A 
/O 


TJi^j-Ti 1 1 in 
dU+Zj J 1 JU 


— Z.oU 


jZoj 


01 

Z.oj 




1 A 

l.o 


n n7 
— U.U / 


n A/1 
— U.04 


1 ^n 
1 jU 


—U.U/ 


n 

— u.jy 


1 TJ 
1-5 J 


n n/1 
— U.U4 


A QA 

—U.JO 


1 n7 
lU / 


lO/D 


— z . ou 


AO /I 1 
0Z4i 


1 SI 
J.oi 




1 A 
1.0 


n ni 
— U.Ul 


n 1 
— U.Zl 


1 1 Q 

1 ly 


A AI 
—U.Ul 


n 1 /I 
— U. 14 


1 no 

lU/ 


n AA 

U.UU 


A AO 
—U.UZ 


88 
oo 


DO ioU / /-UU / 


—Z.oU 


Dj44 


A OQ 

4.zy 




1 A 
1.0 


—U.U J 


— U.jO 


/ / 


_A AA 
—U.UO 


A '\A 
— U.J4 


A7 
0/ 


n A'? 

—U.U J 


_A 00 

—U.zy 


'^0 
jZ 


DO 10UoU-UJ'+ 


— J.UU 


Aitf\< 
4oUj 


1 Ait 




1 
Z. 1 


—U.U J 


— u.zy 


01 

y 1 


—A AO 
—U.UZ 


—U.U J 


SA 
oU 


A Al 
—U.Ul 


_A A^ 
— U.Uj 




DO loucsu-uyj 


— J.ZU 


4y4U 


1 QA 

l.oO 




1 o 

i.y 


n no 
— u.uy 


n AQ 
— u.oy 


Q/1 

y4 


_A A7 
—U.U/ 


n ^0 
— U.DZ 


7Q 
/o 


n (\A 
— U.U4 


_A 01 


Al 
01 


DO lOUo4-loU 


—J. ID 


4/Zo 


1 OA 




O 1 
Z.l 


n no 

—U.uy 


n <Q 

— u.Dy 


1 


_n n7 
—U.U/ 


n /K 
— U.4j 


1 "27 
1 J / 


n ni 
—U.U J 


_A 1 7 

— U.l / 


1 AC 

IUj 


xSo lOD4D-U(5y 


■3 cn 


ODUU 


/t o< 
4.ZD 




1 A 

1.0 


n f\i 
—U.U J 


n "in 
— U.oU 


07 
Zl 


_n ni 
—U.Ul 


n A'j 
—U.U J 


Ol 
Zl 


n lA 
— U.IU 


_A cn 
— U.jU 


1 A 

10 


DO 10DjU-Uo / 


—j.jk) 


4/jU 


l.jl 






Z.J 


nil 

— U.ll 


— U.Oj 


1 00 
IZZ 


_A AQ 
—U.UO 


n no 
—u.uy 


1 

iuj 


A A^ 
—U.U J 


_A OA 
— U.ZO 


7Q 
/o 


rSo lOyZo-yjjj 


— j.Uj 


4oy 1 


1.10 




1 

Z. 1 


— u.uo 


n Ai 

— U.4/ 


AQ 

oy 


_A AA 
—U.UO 


— U.44 


Al 
01 


A AA 
—U.UO 


_A ■^Q 

— u.jy 


40 




—J. J J 


jZ4j 


7n 
Z. /U 




1 A 
1.0 


n HA 
—U.UO 


n <A 

— U.JD 


Q7 

y / 


n A< 

—U.U J 


n /1 1 

— U.41 


oJ 


n no 
—U.UZ 


A on 
— U.ZU 


Al 


13 C 1 A/1 AT HAT 


"J 7n 
—J. /U 


^ ICQ 


J.U4 




1 7 

1. / 


n in 
— U.IU 


n A7 
— U.O / 


<0 

jZ 


n no 
—U.UZ 


1 A 

— U. 1 o 


/I 1 
41 


n ni 
—U.Ul 


n A^ 
—U.U J 


1A 
J 4 


L-o 2/5 /2-lUZ 


o on 

— z.yu 


jyo4 


Q Al 




1 c 
l.D 


n nn 
U.UU 


n no 
— U.UZ 


AC 
Do 


U.UU 


n n^ 
—U.U J 


^7 


n ni 
U.Ul 


n nA 
U.UO 


/I Q 

4o 


/"■Q T>S7S ("107 
ZZo / o-UZ / 


9 ^n 
— z. ju 


DJ4o 


A '^0 

4. jy 




1 A 
1 .4 


—U.U J 


(\ "nO 
— U.jZ 


AI 
Dl 


A A'n 
—U.U J 


n TO 
— U. jv 


J 1 


n no 

—U.UZ 


— U. 1 J 


A^ 

41 




— J.UU 


J loo 




Z. Jj 




1 c 
l.J 


n 1 n 
— U. lU 


n on 
— U. /U 


7C 

/o 


A AO 

—U.uy 


n <Q 
— U.Do 


AQ 

Oo 


n no 

—U.uy 


A ^ A 

— U. jO 


J J 


V.I i. o c3 ivycj 


—3.15 


6290 


4 44 




1.4 


-0.05 


-0.33 


39 


0.00 


0.01 


31 


-0.06 


—0.33 


24 


CS 22944-032 


-3.00 


5300 


2.87 




1.5 


-0.04 


-0.46 


124 


-0.04 


-0.39 


112 


-0.01 


-0.09 


86 


CS 22957-022 


-2.90 


5163 


2.47 




1.6 


-0.07 


-0.58 


125 


-0.06 


-0.49 


114 


-0.03 


-0.29 


88 


CS 22963-004 


-3.50 


5659 


3.40 




1.6 


-0.09 


-0.67 


60 


-0.07 


-0.45 


49 


-0.05 


-0.30 


36 


CS 22965-054 


-3.10 


6205 


3.73 




1.7 


-0.01 


-0.11 


44 


0.03 


0.23 


36 


0.00 


0.01 


30 


CS 29502-092 


-3.30 


4890 


1.72 




1.6 


-0.01 


-0.11 


142 


0.00 


-0.03 


125 


0.01 


0.13 


96 


CS 29506-007 


-2.85 


6369 


3.84 




1.4 


-0.02 


-0.25 


59 


-0.03 


-0.22 


50 


-0.04 


-0.27 


39 


CS 29522-046 


-2.10 


6055 


3.80 




1.4 


-0.02 


-0.27 


130 


-0.02 


-0.23 


117 


0.00 


-0.02 


94 


CS 30312-059 


-3.15 


5021 


2.06 




1.8 


-0.05 


-0.45 


90 


-0.04 


-0.29 


80 


-0.02 


-0.16 


65 


CS 30325-028 


-2.90 


4911 


1.70 




1.8 


-0.07 


-0.60 


122 


-0.06 


-0.50 


110 


-0.03 


-0.22 


84 


CS 30336-049 


-3.95 


4827 


1.51 




2.3 


-0.13 


-0.65 


78 


-0.11 


-0.45 


66 


-0.01 


-0.06 


46 


CS 31078-018 


-2.85 


5257 


2.75 




1.5 


-0.07 


-0.65 


125 


-0.07 


-0.60 


110 


-0.03 


-0.37 


86 


CS 31085-024 


-2.70 


5949 


4.57 




0.9 


-0.07 


-0.63 


101 


-0.06 


-0.52 


89 


-0.04 


-0.36 


69 



TABLE 6 

Comparison of Atmospheric Parameters 



17 





This Study 




Previous Study 






Star ID 


Tcff 


log? 


V, 


Tcff 


log g V, 


Ref. 


BD+03 740 


6557 


4.28 


1.8 


6330 


3.55 


1.4 


1 


BD+24 1676 


6241 


3.81 


1.6 


6250 


3.45 


1.2 


1 


BS 16080-054 


4805 


1.48 


2.1 


4800 


1.1 


2.4 


2 


BS 16084-160 


4728 


1.26 


2.1 


4650 


1.1 


2.2 


2 


CS 30312-059 


5021 


2.06 


1.8 


4950 


2.0 


1.8 


2 


CS 30325-028 


4911 


1.70 


1.8 


4900 


1.8 


2.0 


2 


BS 16467-062 


5388 


3.04 


1.7 


5200 


2.5 


1.6 


3 


BS 16467-062 


5388 


3.04 


1.7 


5364 


2.95 


1.6 


4 



Ref ere nces. — ._, 

(I2005h : (3) ICavrel et all (l2004h : (4) ICohen et alJ 1 I2OO; 



TABLE 7 

Atmospheric errors for BD+03 740 



Element 


Tcff 


log? 


V, 


'^log. 






-1-100 K 


-hO.2 dex 


-1-0.2 km s"' 






Fel 


0.08 


0.00 


-0.02 


0.08 


0.00 


Fell 


0.02 


0.07 


0.00 


0.05 


0.00 


Lil 


0.07 


0.00 


0.07 


0.10 


0.09 


C 


0.13 


0.00 


-0.05 


0.14 


0.06 


Na' 


0.07 


-0.01 


0.07 


0.11 


0.09 


Mg 


0.05 


-0.02 


-0.02 


0.07 


0.01 


Al 


0.07 


0.00 


-0.01 


0.07 


0.01 


Si 


0.07 


-0.02 


-0.06 


0.10 


0.03 


Ca 


0.06 


0.00 


-0.01 


0.06 


0.02 


Sc II 


0.05 


0.07 


-0.01 


0.03 


0.03 


Til 


0.08 


0.01 


0.00 


0.07 


0.02 


Tin 


0.05 


0.05 


-0.03 


0.05 


0.05 


VI 


0.08 


-0.01 


-0.04 


0.10 


0.02 


VII 


0.05 


0.05 


0.00 


0.02 


0.08 


Cri 


0.10 


0.00 


-0.01 


0.10 


0.02 


Cm 


0.04 


0.05 


-0.03 


0.05 


0.04 


Mnl 


0.09 


0.00 


-0.01 


0.09 


0.01 


Mn II 


0.04 


0.05 


-0.01 


0.02 


0.04 


Co 


0.09 


0.01 


0.00 


0.08 


0.01 


Ni 


0.08 


0.00 


-0.01 


0.08 


0.01 


Cu 


0.11 


0.00 


0.00 


0.11 


0.04 


SrII 


0.06 


0.06 


-0.05 


0.07 


0.07 


YII 


0.05 


0.06 


0.00 


0.01 


0.05 


Zrll 


0.06 


0.05 


-0.01 


0.02 


0.06 


Ball 


0.07 


0.06 


0.00 


0.02 


0.07 



' EWs from CS 22878-027 



18 



TABLE 8 

Atmospheric ERRORS forCS 31078-018 



Element 


Tgff 


log g 


V/ 


lOge 


[A. /re] 






•Hj.z aex 


_i.n Vm o~l 






Pel 


0.10 


-0.02 


-0.05 


0.03 


0.00 


Fen 


0.02 


0.07 


-0.02 


0.10 


0.00 


C 


0.20 


-0.05 


0.00 


0.12 


0.11 


N 


0.20 


-0.05 


0.00 


0.12 


0.11 


O 


0.20 


-0.10 


0.00 


0.09 


0.14 


Na' 


0.13 


-0.05 


-0.07 


0.08 


0.11 


Me 


0.09 


-0.04 


-0.02 


0.05 


0.09 


Al 


0.11 


-0.03 


-0.07 


0.04 


0.05 


Si2 


0.13 


—0 07 


—0 05 


1 1 


15 


Ca 


0.07 


00 


—0.01 


07 


0.03 


Sc II 


0.05 


05 


—0 03 


Oil 


03 


Ti I 


Oil 


A on 


—0 01 


Oil 


05 


Ti II 


0.06 


0.05 


—0 04 


13 


03 


V I 


Oil 


00 


00 


Oil 


06 


V II 


06 


Oft 


—0 01 


13 


07 


Cr I 


12 


—0.01 


—0.06 


08 


03 


Cr II 


04 




—0 1 1 


13 


OS 


Mn I 


1 1 


—0 02 


_() oj^ 


05 


03 


Mn II 


05 


04 


—0 08 


12 


0.03 




12 


u.uu 


—A OS 


n in 




Ni 


0.11 


-0.01 


-0.08 


0.08 


0.04 


Cu 


0.14 


-0.01 


-0.11 


0.11 


0.07 


Zn 


0.04 


0.03 


0.00 


0.08 


0.07 


SrH 


0.02 


0.06 


0.00 


0.09 


0.03 


Y n 


0.07 


0.06 


-0.03 


0.15 


0.03 


Zr n 


0.07 


0.06 


—0.02 


0.14 


0.03 


Ru 


13 


00 


—0 02 


12 


07 


Pd 


0.12 


0.00 


0.00 


0.12 


0.07 


Ball 


0.12 


0.01 


-0.16 


0.16 


0.08 


La II 


0.07 


0.07 


-0.01 


0.16 


0.06 


Cell 


0.07 


0.07 


0.00 


0.15 


0.06 


Ndll 


0.07 


0.07 


0.00 


0.15 


0.06 


Smll 


0.07 


0.07 


0.00 


0.15 


0.06 


EuII 


0.08 


0.07 


-0.02 


0.17 


0.03 


Gd II 


0.07 


0.06 


0.00 


0.14 


0.05 


Dy II 


0.08 


0.06 


-0.02 


0.15 


0.04 


Hon 


0.07 


0.07 


-0.01 


0.16 


0.06 


Ern 


0.08 


0.07 


0.00 


0.17 


0.07 


Ybll 


0.10 


0.02 


-0.17 


0.18 


0.02 



' EW from BD+23 3130^ EW from BS 16929-005 



19 



TABLE 9 

Atmospheric errors for CS 29502-092 



Element 


Tgff 


log g 


V/ 


lOge 






.inn ir 


uex 


_i.n Vm o~l 






Fel 


0.12 


-0.03 


-0.06 


0.09 


0.00 


Fen 


0.01 


0.06 


-0.02 


0.08 


0.00 


C 


0.25 


-0.10 


0.00 


0.09 


0.08 


N 


0.30 


0.00 


0.00 


0.30 


0.22 


O 


0.25 


-0.10 


0.00 


0.09 


0.08 


Mg 


0.11 


-0.05 


-0.04 


0.01 


0.04 


Na' 


0.14 


-0.08 


-0.09 


0.05 


0.06 


Al 


0.12 


-0.04 


-0.09 


0.09 


0.03 


Si' 


0.18 


—0.08 


—0.04 


0.02 


0.01 


Ca 


07 


—0.01 


—0.01 


06 


0.04 


Sc 11 


\J.\JO 


u.uu 






U.UJ 


Ti 1 


0.14 


—0.01 


—0.01 


0.13 


0.06 


Ti II 


0.05 


0.06 


—0.03 


0.13 


0.04 


V I 


14 


—0.01 


—0.01 


13 


0.06 


V II 


06 


05 


—0 02 


12 


0.06 


Cr I 


13 


—0 02 


_() 0^ 


10 


03 


Cr II 


—0 02 


06 


00 


03 


04 


Mn I 


0.12 


-0.01 


-0.05 


0.12 


0.02 


Mn n 


0.04 


0.04 


-0.08 


0.13 


0.05 


Co 


0.14 


-0.02 


-0.08 


0.13 


0.04 


Ni 


0.11 


-0.02 


-0.08 


0.11 


0.02 


Cu 


0.17 


-0.03 


-0.10 


0.15 


0.06 


Zn 


0.05 


0.03 


0.00 


0.09 


0.05 


Srn 


0.08 


0.03 


-0.19 


0.24 


0.16 


Yn 


0.08 


0.06 


-0.01 


0.15 


0.07 


Zrn 


0.08 


0.06 


-0.02 


0.16 


0.07 


Pd 


0.17 


-0.03 


-0.10 


0.15 


0.06 


Cd 


0.18 


0.01 


0.00 


0.19 


0.11 


Ball 


0.08 


0.06 


-0.02 


0.16 


0.07 


Latf 


0.08 


0.06 


-0.01 


0.15 


0.07 


Eutf 


0.08 


0.06 


-0.01 


0.15 


0.04 



1 EWs from BS 16080-054^ EWs from CS 30325-028 



TABLE 10 

Abundances relative to Fe (and Fe II for ionized species) 



Star 


[Fe/H] 


numb. 


a 


total 


rijA „/TTi 
[re D/rlJ 


numb. 




total 


[C/reJ 


numb. 




total 


[N/Fe] 


numb. 




total 


[O/Fe] numb. 


(J 


total 


lU 




lines 


lines 


error 




lines 


lines 


error 




lines 


lines 


error 




lines 


lines 


error 


lines 


lines 


error 


BD+03 740 


-2.71 


102 


0.10 


0.08 


-2.69 


10 


0.10 


0.06 


0.59 




0.30 


0.31 


<0.29 








<0.98 


1 






BD+23 3130 


-2.58 


150 


0.13 


0.03 


-2.62 


21 


0.16 


0.11 


0.11 




0.20 


0.23 


-0.54 




0.20 


0.23 


0.55 


1 


0.30 


0.33 


BD+24 1676 


-2.63 


119 


0.07 


0.08 


-2.60 


14 


0.09 


0.06 


0.41 




0.20 


0.21 


<0.21 








0.65 


1 


0.30 




BS 16077-007 


-2.72 


77 


0.10 


0.08 


-2.82 


4 


0.14 


0.09 


0.60 




0.30 


0.31 


<0.80 








<1.39 


1 






BS 16080-054 


-3.07 


91 


0.12 


0.09 


-2.94 


11 


0.10 


0.09 


-0.45 




0.20 


0.22 


0.75 




0.30 


0.37 











BS 16080^093 


-3.19 


94 


0.14 


0.09 


-3.11 


12 


0.14 


0.09 


< -0.63 

























BS 16084-160 


-3.20 


155 


0.17 


0.09 


-3.26 


16 


0.11 


0.08 


-0.12 




0.20 


0.22 


0.78 




0.20 


0.30 


0.27 


1 


0.30 


0.31 


BS 16467-062 


-3.75 


52 


0.15 


0.04 


-3.75 


3 


0.16 


0.14 


0.48 




0.20 


0.23 



















BS 16545-089 


-3.44 


27 


0.06 


0.08 


-3.44 


27 


0.06 


0.08 


<1.62 








<1.82 

















BS 16550-087 


-3.53 


122 


0.17 


0.09 


-3.51 


12 


0.12 


0.09 


-0.49 




0.20 


0.22 


1.11 




0.30 


0.37 











BS 16928-053 


-3.07 


69 


0.15 


0.09 


-3.11 


4 


0.04 


0.08 


-0.25 




0.20 


0.22 


1.05 




0.30 


0.37 











BS 16929-005 


-3.34 


97 


0.12 


0.03 


-3.39 


7 


0.12 


0.11 


0.97 




0.20 


0.23 


0.32 




0.30 


0.32 











CS 22872-102 


-2.87 


68 


0.07 


0.03 


-2.98 


4 


0.14 


0.12 


0.60 




0.20 


0.23 


<0.55 

















CS 22878-027 


-2.48 


61 


0.11 


0.08 


-2.68 


2 


0.08 


0.08 


0.86 




0.10 


0.12 


<1.06 

















CS 22880-086 


-3.01 


78 


0.13 


0.03 


-3.03 


9 


0.14 


0.11 


0.24 




0.15 


0.19 



















CS 22884-108 


-3.14 


39 


0.12 


0.08 


-3.32 


2 


0.06 


0.07 


<0.92 








<1.22 

















CS 22944-032 


-2.98 


124 


0.11 


0.03 


-3.04 


10 


0.09 


0.10 


0.31 




0.10 


0.15 


< -0.44 








0.75 


1 


0.30 


0.33 


CS 22957-022 


-2.93 


125 


0.13 


0.03 


-2.92 


12 


0.10 


0.10 


0.16 




0.20 


0.23 


0.21 




0.20 


0.23 


0.60 


1 


0.30 


0.33 


CS 22963-004 


-3.42 


60 


0.12 


0.03 


-3.44 


3 


0.15 


0.13 


0.40 




0.20 


0.23 


0.80 




0.20 


0.23 


0.99 


1 


0.30 


0.33 


CS 22965-054 


-3.09 


44 


0.11 


0.03 


-3.05 


4 


0.20 


0.14 


<1.07 

























CS 29502-092 


-3.18 


142 


0.09 


0.09 


-3.20 


17 


0.12 


0.09 


0.96 




0.20 


0.22 


0.81 




0.15 


0.27 


0.75 


1 


0.30 


0.31 


CS 29506-007 


-2.85 


59 


0.09 


0.08 


-2.89 


4 


0.09 


0.07 


<0.83 








<0.93 








<1.52 


1 






CS 29522-046 


-2.09 


130 


0.07 


0.08 


-2.24 


18 


0.12 


0.06 


0.42 




0.20 


0.21 


-0.33 




0.30 




0.66 


1 


0.20 




CS 30312-059 


-3.14 


90 


0.12 


0.09 


-3.19 


8 


0.05 


0.08 


0.27 




0.10 


0.13 


< -0.48 

















CS 30325-028 


-2.90 


122 


0.14 


0.09 


-2.96 


16 


0.11 


0.08 


0.38 




0.20 


0.22 


-0.22 




0.15 


0.27 


0.57 


1 


0.30 


0.31 


CS 30336-049 


-4.04 


78 


0.18 


0.09 


^.16 


4 


0.10 


0.09 


-0.28 




0.30 


0.31 


0.92 




0.20 


0.30 


1.01 


1 


0.30 


0.31 


CS 31078-018 


-2.84 


125 


0.12 


0.03 


-2.91 


10 


0.13 


0.11 


0.37 




0.15 


0.19 


-0.38 




0.20 


0.23 


0.81 


1 


0.30 


0.33 


CS 31085-024 


-2.68 


101 


0.12 


0.08 


-2.71 


5 


0.13 


0.08 


0.36 




0.20 


0.21 


<-0.24 








0.65 


1 


0.30 





TABLE 1 1 

Abundances relative to Fe (and Fe II for ionized species) 



Star 


[JNa/reJ 


numb. 


(J 


total 


[Mg/FeJ 


numb. 




total 


[Al/Fe] 


numb. 


<7 


total 


[Sl/reJ 


numb. 


(7 


total 


[Ca/Fe] 


numb. 




total 


lU 




lines 


lines 


error 




lines 


lines 


error 




lines 


lines 


error 




lines 


lines 


error 




lines 


lines 


error 


BD+03 740 











0.23 


6 


0.04 


0.02 


-0.84 


1 


0.15 


0.15 


0.07 


1 


0.15 


0.15 


0.38 




0.04 


0.03 


BD+23 3130 


-0.09 


2 


0.08 


0.12 


0.25 


7 


0.09 


0.10 


-0.99 


2 


0.04 


0.06 











0.24 


14 


0.09 


0.04 


BD+24 1676 











0.37 


6 


0.07 


0.03 











0.17 


1 


0.15 


0.15 


0.45 


14 


0.08 


0.03 


BS 16077-007 











0.34 


5 


0.08 


0.04 











0.13 


1 


0.15 


0.15 


0.43 


10 


0.03 


0.02 


BS 16080-054 


0.48 


2 


0.15 


0.12 


0.44 


5 


0.09 


0.06 


-0.70 


2 


0.04 


0.04 


0.47 


1 


0.15 


0.15 


0.41 


11 


0.06 


0.05 


BS 16080-093 











0.44 


8 


0.07 


0.05 


-0.88 


2 


0.07 


0.06 


0.42 


1 


0.15 


0.15 


0.27 


8 


0.09 


0.05 


BS 16084-160 


-0.22 


2 


0.11 


0.10 


0.21 


7 


0.07 


0.05 


-0.97 


2 


0.04 


0.05 


0.62 


1 


0.15 


0.15 


0.18 


10 


0.10 


0.05 


BS 16467-062 











0.23 


4 


0.10 


0.11 


-0.79 


2 


0.04 


0.06 


0.32 


1 


0.15 


0.21 


0.28 


4 


0.09 


0.06 


BS 16545-089 


-0.13 


1 


0.15 


0.18 


<0.29 


1 






-0.66 


1 


0.15 


0.15 


0.00 


1 


0.15 


0.15 


0.46 


1 


0.15 


0.15 


BS 16550-087 


0.01 


2 


0.08 


0.08 


0.58 


9 


0.06 


0.05 


-0.61 


2 


0.04 


0.05 


0.40 


1 


0.15 


0.15 


0.23 


9 


0.13 


0.06 


BS 16928-053 


0.03 


2 


0.08 


0.08 


0.29 


5 


0.11 


0.07 


-0.90 


2 


0.04 


0.05 


1.16 


1 


0.15 


0.15 


0.21 


11 


0.08 


0.05 


BS 16929-005 


-0.01 


2 


0.08 


0.12 


0.45 


7 


0.09 


0.10 


-0.66 


2 


0.04 


0.06 











0.33 


5 


0.06 


0.04 


CS 22872-102 


-0.04 


2 


0.08 


0.12 


0.38 


3 


0.09 


0.10 


-0.72 


2 


0.11 


0.09 


0.28 


1 


0.15 


0.21 


0.40 


13 


0.05 


0.03 


CS 22878-027 


-0.28 


2 


0.17 


0.15 


0.02 


2 


0.15 


0.11 


-0.84 


1 


0.15 


0.15 


0.07 


1 


0.15 


0.15 


0.18 


7 


0.18 


0.07 


CS 22880-086 











0.29 


4 


0.08 


0.10 


-0.97 


2 


0.08 


0.08 











0.26 


4 


0.11 


0.06 


CS 22884-108 











<0.18 


1 






-0.73 


2 


0.07 


0.05 


0.11 


1 


0.15 


0.15 


0.36 


3 


0.08 


0.06 


CS 22944-032 











0.39 


8 


0.07 


0.09 


-0.84 


1 


0.15 


0.16 











0.29 


10 


0.06 


0.04 


CS 22957-022 











0.28 


8 


0.09 


0.10 




















0.22 


10 


0.07 


0.04 


CS 22963-004 











0.35 


5 


0.06 


0.10 


-0.85 


1 


0.15 


0.16 


0.42 


1 


0.15 


0.21 


0.31 


3 


0.08 


0.06 


CS 22965-054 











0.31 


2 


0.09 


0.11 


-0.73 


1 


0.15 


0.16 


0.08 


1 


0.15 


0.21 


0.43 


3 


0.11 


0.07 


CS 29502-092 











0.42 


7 


0.09 


0.05 


-0.79 


1 


0.15 


0.15 











0.31 


10 


0.10 


0.05 


CS 29506-007 











0.28 


4 


0.10 


0.05 


-0.82 


1 


0.15 


0.15 


0.10 


1 


0.15 


0.15 


0.45 


5 


0.09 


0.05 


CS 29522-046 











0.40 


5 


0.04 


0.02 




















0.38 


14 


0.09 


0.03 


CS 30312-059 


0.17 


2 


0.08 


0.08 


0.41 


5 


0.08 


0.06 


-0.88 


2 


0.07 


0.06 











0.33 


12 


0.06 


0.05 


CS 30325-028 











0.41 


7 


0.04 


0.04 




















0.32 


9 


0.09 


0.05 


CS 30336-049 











0.08 


6 


0.11 


0.06 











0.19 


1 


0.15 


0.15 


0.09 


5 


0.13 


0.07 


CS 31078-018 











0.42 


8 


0.08 


0.09 


-0.73 


1 


0.15 


0.16 











0.41 


11 


0.10 


0.04 


CS 31085-024 











0.17 


7 


0.07 


0.03 











0.21 


1 


0.15 


0.15 


0.22 


10 


0.07 


0.03 



TABLE 12 

Abundances relative to Fe (and Fe II for ionized species) 



Star 

lU 


[ScU/reJ 


numb, 
lines 


(J 
lines 


total 
error 


[Ti/Fe] 


numb, 
lines 


lines 


total 
error 


[li n/rej 


numb, 
lines 


lines 


total 
error 


[V/Fe] 


numb, 
lines 


(J 
lines 


total 
error 


[V ii/rej 


numb, 
lines 


lines 


total 
error 


BD+03 740 


0.29 


4 


0.06 


0.05 


0.69 


7 


0.07 


0.03 


0.57 


28 


0.11 


0.06 











0.28 


4 


0.06 


0.09 


BD+23 3130 


0.06 


6 


0.06 


0.05 


0.21 


19 


0.06 


0.05 


0.25 


38 


0.13 


0.05 


0.08 


3 


0.25 


0.16 


0.12 


8 


0.09 


0.08 


BD+24 1676 


0.20 


5 


0.06 


0.05 


0.55 


9 


0.08 


0.03 


0.45 


32 


0.11 


0.06 











0.20 


5 


0.09 


0.09 


BS 16077-007 


0.31 


3 


0.06 


0.08 


0.59 


3 


0.10 


0.06 


0.62 


23 


0.09 


0.07 











0.42 


1 


0.15 


0.18 


BS 16080-054 


-0.02 


5 


0.16 


0.09 


0.24 


6 


0.10 


0.07 


0.30 


24 


0.10 


0.06 











-0.03 


5 


0.07 


0.07 


BS 16080-093 


0.09 


3 


0.06 


0.07 


0.38 


9 


0.07 


0.07 


0.41 


28 


0.13 


0.06 











-0.09 


1 


0.15 


0.17 


BS 16084-160 


0.09 


6 


0.07 


0.06 


-0.02 


11 


0.08 


0.07 


0.09 


28 


0.09 


0.06 











-0.14 


6 


0.06 


0.07 


BS 16467-062 


0.31 


2 


0.06 


0.11 


0.51 


1 


0.15 


0.16 


0.44 


9 


0.12 


0.10 




















BS 16545-089 


0.29 


1 


0.15 


0.15 











0.71 


10 


0.08 


0.06 




















BS 16550-087 


-0.05 


4 


0.13 


0.09 


0.18 


7 


0.13 


0.08 


0.15 


20 


0.11 


0.07 




















BS 16928-053 


-0.16 


4 


0.11 


0.08 


-0.02 


4 


0.08 


0.07 


0.21 


16 


0.10 


0.09 











-0.37 


1 


0.15 


0.16 


BS 16929-005 


-0.03 


1 


0.15 


0.16 


0.45 


6 


0.05 


0.06 


0.50 


18 


0.13 


0.06 











0.01 


1 


0.15 


0.17 


CS 22872-102 


0.08 


3 


0.06 


0.08 


0.48 


3 


0.08 


0.07 


0.42 


20 


0.13 


0.05 




















CS 22878-027 


0.16 


1 


0.15 


0.16 


0.46 


2 


0.16 


0.12 


0.53 


12 


0.12 


0.10 




















CS 22880-086 


0.06 


2 


0.07 


0.07 


0.23 


4 


0.12 


0.08 


0.36 


13 


0.10 


0.06 




















CS 22884-108 


0.45 


1 


0.15 


0.16 











0.80 


10 


0.12 


0.11 




















CS 22944-032 


0.11 


3 


0.11 


0.08 


0.32 


13 


0.09 


0.06 


0.36 


25 


0.12 


0.05 











0.16 


5 


0.04 


0.08 


CS 22957-022 


0.05 


3 


0.06 


0.06 


0.30 


8 


0.06 


0.06 


0.33 


26 


0.08 


0.05 


-0.09 


1 


0.15 


0.16 


0.04 


5 


0.04 


0.08 


CS 22963-004 


0.27 


2 


0.06 


0.10 











0.44 


17 


0.12 


0.08 











0.07 


1 


0.15 


0.19 


CS 22965-054 


0.03 


1 


0.15 


0.18 











0.42 


10 


0.12 


0.07 




















CS 29502-092 


0.02 


6 


0.04 


0.06 


0.18 


19 


0.08 


0.06 


0.19 


27 


0.12 


0.05 


-0.06 


2 


0.16 


0.13 


0.04 


6 


0.02 


0.07 


CS 29506-007 


0.28 


2 


0.08 


0.08 


0.66 


2 


0.08 


0.06 


0.57 


20 


0.13 


0.07 




















CS 29522-046 


0.12 


5 


0.04 


0.04 


0.35 


14 


0.07 


0.03 


0.43 


32 


0.12 


0.06 


0.06 


1 


0.15 


0.15 


0.17 


6 


0.06 


0.09 


CS 30312-059 


0.07 


3 


0.06 


0.06 


0.31 


5 


0.07 


0.07 


0.39 


17 


0.10 


0.06 











0.04 


5 


0.04 


0.07 


CS 30325-028 


0.08 


5 


0.09 


0.07 


0.26 


19 


0.08 


0.06 


0.35 


32 


0.12 


0.06 


-0.11 


1 


0.15 


0.16 


0.03 


5 


0.05 


0.07 


CS 30336-049 


0.18 


4 


0.07 


0.08 


0.16 


1 


0.15 


0.16 


0.27 


22 


0.12 


0.10 




















CS 31078-018 


0.10 


4 


0.06 


0.06 


0.31 


11 


0.07 


0.06 


0.38 


26 


0.09 


0.05 


0.06 


2 


0.16 


0.13 


0.13 


5 


0.04 


0.08 


CS 31085-024 


0.14 


2 


0.07 


0.08 


0.32 


4 


0.08 


0.05 


0.38 


22 


0.16 


0.08 











0.14 


3 


0.05 


0.10 



TABLE 13 

Abundances relative to Fe (and Fe II for ionized species) 



Star 


[Cr/Fe] 


numb. 




total 


[Cr n/Fe] 


numb. 


a 


total 


[Mn/Fe] 


numb. 


c 


total 


[Mnu/Fe] 


numb. 


c 


total 


[Co/Fe] 


numb. 


c 


total 


ID 




lines 


lines 


error 




lines 


lines 


error 




Unes 


Unes 


error 




Unes 


Unes 


error 




Unes 


Unes 


error 


BD+03 740 


-0.07 


5 


0.05 


0.03 


0.05 


2 


0.10 


0.09 


-0.46 


3 


0.16 


0.09 


-0.19 


3 


0.06 


0.06 


0.41 


3 


0.09 


0.05 


BD+23 3130 


-0.24 


19 


0.10 


0.04 


-0.07 


4 


0.16 


0.10 


-0.57 


10 


0.13 


0.05 


-0.38 


3 


0.07 


0.06 


0.20 


2 


0.16 


0.12 


BD+24 1676 


-0.13 


7 


0.09 


0.04 


-0.09 


3 


0.12 


0.08 


-0.43 


5 


0.21 


0.09 


-0.30 


3 


0.06 


0.06 


0.35 


2 


0.09 


0.06 


BS 16077-007 


-0.15 


6 


0.06 


0.03 


0.01 


2 


0.10 


0.11 


-0.52 


3 


0.16 


0.09 


-0.27 


2 


0.06 


0.09 


0.34 


2 


0.09 


0.06 


BS 16080-054 


-0.43 


7 


0.26 


0.10 


-0.21 


1 


0.15 


0.16 


-0.61 


4 


0.25 


0.13 


-0.59 


1 


0.15 


0.16 


0.08 


3 


0.09 


0.07 


BS 16080-093 


-0.34 


5 


0.13 


0.07 











-0.98 


3 


0.16 


0.10 











0.30 


2 


0.10 


0.08 


BS 16084-160 


-0.42 


10 


0. 16 


0.06 











-0.55 


9 


0.25 


0.09 


—0.44 


3 


0.06 


0.07 


0.09 


3 


0. 1 1 


0.08 


BS 16467-062 


-0.46 


4 


0.11 


0.06 











-0.64 


3 


0.16 


0.10 











0.69 


3 


0.09 


0.07 


BS 16545-089 


-0.15 


1 


0.15 


0.15 






































BS 16550-087 


-0.59 


4 


0.16 


0.09 











-0.96 


3 


0.16 


0.10 











0.22 


3 


0.09 


0.07 


BS 16928-053 


-0.25 


4 


0.11 


0.06 











-0.71 


4 


0.24 


0.12 


-0.50 


2 


0.12 


0.10 


0.04 


3 


0.11 


0.08 


BS 16929-005 


-0.50 


4 


0.11 


0.06 











-0.80 


3 


0.16 


0.10 




2 


0.10 


0.09 


0.38 


3 


0.09 


0.06 


CS 22872-102 


-0.32 


5 


0.08 


0.05 











-0.62 


3 


0.16 


0.10 


-0.44 


3 


0.10 


0.10 


0.22 


3 


0.09 


0.06 


CS 22878-027 


-0.12 


5 


0.06 


0.04 











-0.34 


3 


0.16 


0.09 


-0.04 


1 


0.15 


0.17 











CS 22880-086 


-0.37 


2 


0.11 


0.08 











-0.50 


4 


0.27 


0.14 











0.47 


3 


0.09 


0.07 


CS 22884-108 


-0.25 


4 


0.11 


0.06 











-0.50 


2 


0.16 


0.11 




















CS 22944-032 


-0.31 


9 


0.10 


0.05 











-0.44 


6 


0.21 


0.09 


-0.19 


3 


0.06 


0.05 


0.45 


3 


0.09 


0.06 


CS 22957-022 


-0.30 


9 


0.13 


0.05 











-0.76 


3 


0.16 


0.10 


-0.25 


3 


0.06 


0.05 


0.29 


3 


0.13 


0.09 


CS 22963-004 


-0.40 


5 


0.08 


0.05 











-0.77 


3 


0.16 


0.10 


-0.42 


1 


0.15 


0.18 


0.48 


2 


0.09 


0.07 


CS 22965-054 


-0.05 


3 


0.11 


0.07 











-0.30 


2 


0.16 


0.12 











0.26 


1 


0.15 


0.16 


CS 29502-092 


-0.21 


15 


0.12 


0.04 


0.21 


2 


0.10 


0.08 


-0.32 


8 


0.26 


0.09 


-0.16 


3 


0.06 


0.07 


0.36 


3 


0.09 


0.06 


CS 29506-007 


-0.17 


6 


0.08 


0.04 











-0.69 


3 


0.16 


0.09 


-0.42 


1 


0.15 


0.16 


0.28 


1 


0.15 


0.15 


CS 29522-046 


-0.12 


15 


0.06 


0.03 


0.11 


3 


0.12 


0.08 


-0.47 


7 


0.16 


0.06 


-0.27 


3 


0.06 


0.06 


0.17 


3 


0.09 


0.05 


CS 30312-059 


-0.34 


6 


0.20 


0.09 


0.09 


1 


0.15 


0.16 


-0.88 


3 


0.16 


0.10 


-0.53 


3 


0.06 


0.06 


0.28 


3 


0.09 


0.07 


CS 30325-028 


-0.26 


15 


0.10 


0.04 


-0.01 


2 


0.10 


0.08 


-0.51 


6 


0.28 


0.12 


-0.41 


3 


0.17 


0.11 


0.23 


3 


0.12 


0.08 


CS 30336-049 


-0.74 


5 


0.16 


0.08 











-0.82 


3 


0.16 


0.10 


-0.52 


2 


0.07 


0.09 


0.47 


3 


0.10 


0.07 


CS 31078-018 


-0.31 


6 


0.11 


0.06 


-0.18 


1 


0.15 


0.16 


-0.20 


6 


0.20 


0.09 


0.01 


3 


0.06 


0.06 


0.34 


3 


0.16 


0.10 


CS 31085-024 


-0.23 


7 


0.07 


0.04 


-0.10 


1 


0.15 


0.17 


-0.41 


5 


0.23 


0.10 


-0.19 


3 


0.06 


0.08 


0.37 


3 


0.09 


0.05 



TABLE 14 

Abundances relative to Fe (and Fe II for ionized species) 



Star 


[JNl/reJ 


numb. 


(J 


total 


[Cu/Fe] 


numb. 




total 


[Zn/Fe] 


numb. 




total 


[or ii/rej 


numb. 




total 


[Y D/Fe] 


numb. 


a 


total 


ID 




lines 


lines 


error 




lines 


lines 


error 




lines 


lines 


error 




lines 


lines 


error 




lines 


lines 


error 


BD+03 740 


0.06 


3 


0.08 


0.05 


-0.80 


2 


0.06 


0.06 











-0.21 


2 


0.06 


0.09 


-0.01 


1 


0.15 


0.16 


BD+23 3130 


-0.15 


5 


0.10 


0.06 


-1.33 


2 


0.14 


0.12 


0.07 


2 


0.04 


0.08 


-0.38 


2 


0.10 


0.08 


-0.39 


11 


0.08 


0.05 


BD+24 1676 


0.03 


4 


0.08 


0.04 


-0.83 


2 


0.06 


0.06 


0.20 


1 


0.15 


0.18 


-0.05 


2 


0.05 


0.08 


-0.12 


5 


0.05 


0.06 


BS 16077-007 


0.08 


3 


0.08 


0.05 


-0.59 


1 


0.15 


0.16 











0.22 


2 


0.05 


0.11 


0.30 


1 


0.15 


0.17 


BS 16080-054 


-0.05 


3 


0.11 


0.07 




















0.16 


2 


0.05 


0.17 


0.12 


8 


0.14 


0.09 


BS 16080-093 


0.03 


2 


0.14 


0.10 




















-0.13 


2 


0.05 


0.17 


-0.28 


4 


0.09 


0.09 


BS 16084-160 


-0.13 


4 


0.13 


0.07 


-1.64 


2 


0.11 


0.10 











-2.10 


2 


0.11 


0.18 











BS 16467-062 


0.29 


2 


0.08 


0.07 






































BS 16545-089 





























-0.09 


2 


0.07 


0.09 











BS 16550-087 


-0.08 


3 


0.10 


0.06 




















0.42 


2 


0.05 


0.17 


0.29 


9 


0.09 


0.08 


BS 16928-053 


-0.15 


2 


0.08 


0.06 




















-0.34 


2 


0.05 


0.17 


-0.41 


5 


0.12 


0.09 


BS 16929-005 


0.08 


3 


0.09 


0.07 




















0.37 


2 


0.08 


0.08 


0.22 


4 


0.09 


0.07 


CS 22872-102 


0.03 


3 


0.10 


0.07 




















-0.15 


2 


0.05 


0.09 











CS 22878-027 


0.06 


3 


0.08 


0.05 




















-0.18 


2 


0.05 


0.10 











CS 22880-086 


0.16 


3 


0.08 


0.06 




















-0.10 


2 


0.05 


0.07 


-0.27 


3 


0.15 


0.10 


CS 22884-108 


0.00 


2 


0.11 


0.08 




















0.29 


2 


0.10 


0.11 











CS 22944-032 


0.10 


4 


0.08 


0.06 


-1.10 


2 


0.06 


0.08 


0.25 


1 


0.15 


0.17 


-0.27 


2 


0.10 


0.08 


-0.30 


7 


0.07 


0.05 


CS 22957-022 


-0.02 


3 


0.11 


0.08 


-1.23 


2 


0.07 


0.09 


0.31 


2 


0.04 


0.08 


-0.41 


1 


0.15 


0.16 


-0.37 


6 


0.11 


0.06 


CS 22963-004 


0.08 


2 


0.08 


0.07 


-0.94 


2 


0.07 


0.09 











-0.80 


2 


0.05 


0.10 











CS 22965-054 


0.00 


2 


0.08 


0.07 




















0.06 


2 


0.05 


0.11 











CS 29502-092 


0.08 


5 


0.07 


0.04 


-1.36 


2 


0.11 


0.10 


0.34 


2 


0.04 


0.06 


-0.46 


2 


0.07 


0.17 


-0.63 


8 


0.06 


0.08 


CS 29506-007 


-0.06 


2 


0.08 


0.06 


< -0.56 


1 















0.13 


2 


0.06 


0.09 


0.24 


1 


0.15 


0.16 


CS 29522-046 


-0.03 


5 


0.10 


0.05 


-0.79 


2 


0.06 


0.06 


0.15 


2 


0.05 


0.11 


0.22 


2 


0.06 


0.09 


0.09 


8 


0.05 


0.06 


CS 30312-059 


-0.05 


3 


0.10 


0.06 




















-0.02 


2 


0.06 


0.17 


-0.01 


6 


0.09 


0.08 


CS 30325-028 


-0.04 


4 


0.09 


0.05 


-1.18 


2 


0.06 


0.07 


0.33 


2 


0.09 


0.08 


0.16 


3 


0.15 


0.18 


0.01 


12 


0.06 


0.08 


CS 30336-049 


-0.05 


2 


0.08 


0.06 


-1.17 


1 


0.15 


0.16 











-1.50 


2 


0.10 


0.18 











CS 31078-018 


0.09 


4 


0.10 


0.06 


-1.02 


2 


0.07 


0.09 


0.41 


1 


0.15 


0.17 


0.24 


1 


0.15 


0.16 


0.24 


10 


0.08 


0.06 


CS 31085-024 


0.21 


3 


0.08 


0.05 


-0.93 


2 


0.06 


0.06 











-0.30 


2 


0.05 


0.10 












TABLE 15 

Abundances relative to Fe (and Fe II for ionized species) 



Star 


[Zr D/reJ 


numb. 


(J 


total 


[Ba ii/rej 


numb. 


(J 


total 


[La D/reJ 


numb. 


(J 


total 


[Lu ii/rej 


numb. 




total 






lines 


lines 


error 




lines 


lines 


error 




lines 


lines 


error 




lines 


lines 


error 


BD+03 740 


0.36 


2 


0.11 


0.10 


-0.37 


1 


0.15 


0.17 











<0.68 








BD+23 3130 


0.00 


10 


0.12 


0.06 


-0.47 


3 


0.06 


0.09 


-0.26 


1 


0.15 


0.17 


0.31 


I 


0.10 


0.11 


BD+24 1676 


0.31 


5 


0.14 


0.09 


-0.33 


2 


0.06 


0.09 











<0.59 


I 






BS 16077-007 


0.85 


2 


0.19 


0.16 


-0.11 


1 


0.15 


0.18 











<0.81 


I 






BS 16080-054 


0.31 


9 


0.19 


0.10 


-0.42 


4 


0.06 


0.08 











<0.23 








BS 16080-093 


0.17 


1 


0.15 


0.17 


-0.27 


2 


0.06 


0.09 











0.20 


I 


0.20 


0.21 


BS 16084-160 











-1.99 


2 


0.09 


0.10 











< -0.25 


I 






BS 16467-062 











< -0.68 


1 















<1.24 








BS 16545-089 











<-0.09 


1 















<1.93 


I 






BS 16550-087 


0.74 


8 


0.16 


0.10 


-0.74 


4 


0.15 


0.11 











<0.00 


I 






BS 16928-053 


-0.10 


2 


0.14 


0.12 


-0.82 


4 


0.10 


0.09 











<-0.40 








BS 16929-005 


0.53 


2 


0.11 


0.10 


-0.28 


3 


0.06 


0.10 











<0.03 








CS 22872-102 











-0.45 


1 


0.15 


0.18 











<1.07 


I 






CS 22878-027 











<-0.75 


1 















<1.17 


I 






CS 22880-086 











-0.80 


2 


0.06 


0.10 











<0.22 


] 






CS 22884-108 











0.24 


1 


0.15 


0.17 











<1.91 








CS 22944-032 


0.05 


5 


0.06 


0.05 


-0.62 


2 


0.06 


0.09 











<0.73 








CS 22957-022 


-0.05 


3 


0.11 


0.08 


-1.01 


2 


0.06 


0.10 











< -0.09 








CS 22963-004 











-0.44 


1 


0.15 


0.19 











<1.43 








CS 22965-054 











< -0.48 


1 















<1.14 








CS 29502-092 


-0.28 


5 


0.05 


0.08 


-1.26 


2 


0.06 


0.09 











<-0.31 








CS 29506-007 


0.63 


3 


0.11 


0.10 


0.11 


1 


0.15 


0.17 











<0.88 








CS 29522-046 


0.41 


9 


0.11 


0.08 


0.14 


3 


0.06 


0.08 











<0.23 








CS 30312-059 


0.35 


7 


0.11 


0.08 


-0.01 


4 


0.06 


0.08 











0.63 




0.10 


0.11 


CS 30325-028 


0.40 


19 


0.11 


0.08 


-0.34 


3 


0.06 


0.08 


-0.18 


1 


0.15 


0.17 


0.15 




0.20 


0.21 


CS 30336-049 











-1.32 


1 


0.15 


0.17 











<0.95 








CS 31078-018 


0.63 


16 


0.09 


0.06 


0.72 


3 


0.27 


0.18 


0.74 


1 


0.15 


0.17 


1.23 


4 


0.04 


0.05 


CS 31085-024 











-0.52 


1 


0.15 


0.18 











<0.60 


1 







26 



TABLE 16 
Neutron-capture ABUNDANCES of CS 
31078-018 



Atomic 


Element 


[X/Fe] 


loge(X) 


en'or 


Number 


Name 








38 


Sr II 


31 


30 


18 


39 


Y II 


23 


—0 43 


15 


40 


Zr II 


0.62 


0.32 


0.14 


44 


Rti I 


65 


—0 35 


19 


46 


Pd I 


0.79 


-0.36 


0.19 


56 


Ball 


0.72 


-0.06 


0.31 


57 


La II 


0.74 


-1.00 


0.22 


58 


Cell 


0.66 


-0.67 


0.21 


60 


Ndn 


0.79 


-0.62 


0.21 


62 


SmU 


1.13 


-0.77 


0.16 


63 


Eu II 


1.23 


-1.17 


0.17 


64 


Gdll 


1.03 


-0.76 


0.17 


66 


Dy II 


1.00 


-0.77 


0.15 


67 


Ho II 


1.54 


-1.11 


0.22 


68 


ErII 


1.02 


-0.96 


0.20 


70 


Ybll 


0.76 


-1.07 


0.21 


90 


Thll 


1.47 


-1.35 


0.25 



TABLE 17 

A SAMPLE OF HIGHLY NEUTRON-CAPTURE DEFICIENT 
STARS 



Star 


[Fe/H] 


[Sr/H] 


[Ba/H] 


source 


ID 










BS 16084-160 


-3.26" 


-5.29 


-5.25 


1 


CS 30336-049 


-4.16* 


-5.59 


-5.48 


1 


CS 29502-092 


-3.20* 


-3.59 


-4.46 


1 


HD 4306 


-2.89 


-3.00 


-4.06 


2 


BS 16469-075 


-3.03 


-2.80 


-4.15 


2 


BS 16920-017 


-3.12 


-3.55 


-4.95 


2 


BS 16928-053 


-2.91 


-3.14 


-4.07 


2 


Draco 119 


-2.97 


< -5.47 


< -5.57 


3 


CS 22169-035 


-3.04 


-3.12 


-4.23 


4 


CS 22172-002 


-3.86 


-5.17 


-5.03 


4 


CS 22189-009 


-3.49 


-4.44 


-4.78 


4 


CS 22897-008 


-3.41 


-2.97 


-4.41 


4 


CS 22952-015 


-3.43 


-4.42 


-4.76 


4 


CS 22968-014 


-3.56 


-5.36 


-5.33 


4 


CS 29502-042 


-3.19 


-5.17 


-4.88 


4 


CS 30325-094 


-3.30 


-5.54 


-5.21 


4 


BS 16934-002 


-2.81* 


-4.13 


-4.61 


5 


CS 30327-038 


-2.64* 


-3.28 


-4.14 


5 


HE 1356-0622 


-3.36* 


-5.24 


-4.45 


5 


References. - 


— (1) This 


study; (2) iHonda et al J 


2004D; (3) 



[Fulbright e t all 120041); (4) [ Francois et al J I2007l) ~75) 
12001 ; (6) ICohenet all 120081) 



Aoki et ai] 



[Fell/H] 



27 




50 100 150 200 

EW this study (mA) 




50 100 150 200 

EW this study (mA) 




50 100 150 200 

EW this study (mA) 



Fig. 1 . — The top panel compares the EWs m easured bvlCavrel et al.l 120041) with our study, the middle panel shows the comparison with lAoki et al.l )2005l) . 
and the bottom panel shows the comparison with llvans et 



28 



x; 
-a 



> 

-a 
c 

X 



0.05 
0.00 



-0.15 - 



-0.20 



-o.ioh ° 8 o 



4500 



O A 

o 



5000 



^_ o - - 

A 
O 



o 



O All lines 

A ;^ > 0.3 eV 

O X > 12 eV 



5500 



6000 



6500 



7000 



T (K) 



0.05 

o 0.00^- 
-a 



> 

CD 

-a 
c 

OJ 
-h-' 

X 



-0.15 
-0.20 



-0.05 7 



- 4>- - - 



-4.0 



-3.5 -3.0 
[Fe/H] 



-2.5 



-2.0 



Fig. 2. — Value of the x/loge(Fe I) slopes for each of our stars. When all Fe I lines are considered, there seems to be a trend with both Teff and [Fe/H], although 
considering only the x >1.2 eV lines, this trend largely disappears. Overall, however, even when considering high x lines, we still seem to be finding a negative 
slope in most of our stars. See the electronic edition of the Journal for a color version of this figure. 



m 
o 

o 




6500 6000 5500 5000 4500 
T (K) 

eff ^ ' 



Fig . 3 . — Plot of Teff vs. log g for our stars. The sample spans a wide range of evolutionary states. 



29 



-\ 0.0 
-0.5 
-1.0 

























Aoki et al. 2005 


o 








O 




Ivans et al. 3003 
Cayrel et al. 2004 


A 












Cohen et al. 2008 




■ 






O 












V- 

o 






"a. 




- 




Q 







o 


O 

































10 20 30 40 50 60 
Atomic Number 



Fig. 4. — Comparison of our abundance to that of ottier studies. The sense of the y-axis is the values from this study minus the values from previous studies. 
See the electronic edition of the Journal for a color version of this figure. 



30 




-4.5 -4.0 -3.5 -3.0 -2.5 -2.0 
[Fe/H] 




-4.5 -4.0 -3.5 -3.0 -2.5 -2.0 
[Fe/H] 



Fig. 5. — Values of Loge(Li) and [(C,N,0)/Fe] vs. [Fe/H]. The diamonds (colored green in the electronic edition) are from Spite et al. (2005) and the triangles 
(colored red in the electronic edition) are from Bonifacio et al. (2007). In all of the following figures, the black points with error bars are data from this study. 
See the electronic edition of the Journal for a color version of this figure. 



31 




-1.2 : 
-1.4 T 



-4.5 -4.0 -3.5 -3.0 -2.5 -2.0 
[Fe/H] 



^ 0.5 



-0.5 



-4.5 -4.0 -3.5 -3.0 -2.5 -2.0 
[Fe/H] 



Fig. 6. — [(Na,Al,Sc)/Fe] vs. [Fe/H]. The diamonds (colored green in the electronic edition) are fromlCay rel et alj )2004D and the triangles (colored red in the 
electronic ed ition) are from Cohen et al. (2004). Although we only have two stars below [Fe/H] of -2.7 with measured Na, we do not find the trend found by 
ICavrel et alj (2004). The [Al/Fe] values from Cohen et al^ c2004i) are plotted without the NLTE correction of 0.6 dex assumed in that study. See the electronic 
edition of the Journal for a color version of this figure. 



32 




-4.5 -4.0 -3.5 -3.0 -2.5 -2.0 
[Fe/H] 




-4.5 -4.0 -3.5 -3.0 -2.5 -2.0 
[Fe/H] 



1.0 

0.5 
0.0 
-0.5 



— —1 — — — — li- — — — — — — 



-4.5 -4.0 -3.5 -3.0 -2.5 -2.0 
[Fe/H] 




-4.5 -4.0 -3.5 -3.0 -2.5 -2.0 
[Fe/H] 



Fig. 7. — [a/Fe] vs. [Fe/H]. The symbols are as in figure|6] See the electronic edition of the Journal for a color version of this figure. 



33 




Fig . 8 . — [Si/Fe] vs. [Fe/H] and T^ff. The symbols are as in figure|6] In the bottom panel, we also add the data from lPreston et al] 12003) as the squares (colored 
purple in the electronic edition). See the electronic edition of the Journal for a color version of this figure. 



34 



0.6 : 
0.4 : 




-0.4 : 
-0.6 : 



-3.5 -3.0 -2.5 -2.0 

[Fe/H] 

J_Q= , , ^ 




-3.5 -3.0 -2.5 -2.0 

[Fe/H] 



0.4 - 




-0.2 : 

-0.4 - 



-3.5 -3.0 -2.5 -2.0 

[Fe/H] 

Fig. 9. — Vanadium abundance vs. [Fe/H]. We plot both the neutral and singly ionized species, along with [V IIN I]. We find no offset between the V n and 
V I abundances. 



35 



0.2 : 




-1.0 : 



-4.5 -4.0 -3.5 -3.0 -2.5 -2.0 
[Fe/H] 




-0.4 : : 

-4.5 -4.0 -3.5 -3.0 -2.5 -2.0 



[Fe/H] 



-4.5 -4.0 -3.5 -3.0 -2.5 -2.0 
[Fe/H] 



Fig. 10. — Cr abundance vs. [Fe/H]. We plot both the neutral and singly ionized species, along with [Cr 11/Cr 1]. The symbols are as in figure[6] Although we 
do not measure Cr 11 for many of our stars, our results suggest an offset from zero for [Cr 11/Cr I], as well as an increasing trend in this value with decreasing 
metallicity. See the electronic edition of the Journal for a color version of this figure. 



36 



0.2 
0.0 




-1.2 T 



-4.5 -4.0 -3.5 -3.0 -2.5 -2.0 
[Fe/H] 



0.4 r 

0.2 : 




-4.5 -4.0 -3.5 -3.0 -2.5 -2.0 
[Fe/H] 




-0.2 r 
-0.4 T 



-4.5 -4.0 -3.5 -3.0 -2.5 -2.0 
[Fe/H] 

Fig. 1 1. — Manganese abundance vs. [Fe/H]. We plot both the neutral and sing ly ionized species, along with [Mn II/Mn I]. The symbols are as in figure|6] In 
the [Mn II/Fe] plot, we also overplot the [Mn/Fe] values from lCayrel et alj j2004l) for the reasons described in the text. See the electronic edition of the Journal 
for a color version of this figure. 



37 




Fig. 12. — [Co,Ni/Fe] vs. [Fe/H]. The symbols are as in figure|6] We find a similar trend of increasing [Co/Fe] with decreasing [Fe/H] as in ICavrel et al J 
(20041) . The [Ni/Fe] shows no trend with [Fe/H]. See the electronic edition of the Journal for a color version of this figure. 



El. 



3 
U 




0) 



c 

tsi 



0.6 : 
0.4 : 
0.2 : 
0.0 - 
-0.2 - 



O 0% ' T 



-4.5 -4.0 -3.5 -3.0 -2.5 -2.0 
[Fe/H] 



Fig. 13. — [(Cu,Zn)/Fe] vs. [Fe/H]. The trian gles (colored red in the electronic edition) in the [Cu/Fe] plot are from lCohen et al.l j2008l) . and the diamonds 
(colored green in the electronic edition) are from lBihain et aljj2004l) . In the [Zn/Fe] plot, the symbols are as in figure|6] See the electronic edition of the Journal 
for a color version of this figure. 



38 



~ -1 

m 

-2t- 



^ . * -J o „ 



o o 



-4.5 -4.0 -3.5 -3.0 -2.5 -2.0 
[Fe/H] 



1.0 r 

0.5 - 

J) 

^ 0.0 - 

>1 -0-5 r 
-1.0 - 
-I.5L 



9. , t, 



-4.5 -4.0 -3.5 -3.0 -2.5 -2.0 
[Fe/H] 




-4.5 -4.0 -3.5 -3.0 -2.5 -2.0 
[Fe/H] 




-4.5 -4.0 -3.5 -3.0 -2.5 -2.0 
[Fe/H] 



Fig. 1 4. — [(Sr n, Y II, Zr II, Ba II)/Fe] vs. [Fe/H]. We clearly see a large scatter in all of these abundances. The triangles (colored red in the electronic edition) 
are from lCohen et a l. (2008) and the diamonds (colored green in the electronic edition) are t'rom iFrancois et aU l2007.) . See the electronic edition of the Journal 
for a color version of this figure. 



39 



>- 




-4.0 -3.5 -3.0 -2.5 -3.0 
[Fe/H] 



CO 
[SI 




-4.0 -3.5 -3.0 -2.5 -2.0 
[Fe/H] 




-4.0 -3.5 -3.0 -2.5 -2.0 
[Fe/H] 



Fig. 15. — The light neutron-capture elemen ts, Sr, Zr, and Y, show remarkable correlation, while [Sr/Ba] shows a scatter of almost 2 dex. The diamonds 
(colored green in the electronic edition) are from lFrancois et al] 120071) . See the electronic edition of the Journal for a color version of this figure. 



1.00 



0.95 - 



0.90 T 








4tj 




Ijl 






1 



log 


e 


Th 


= -10.0 


log 


e 


Th 


= -1.55 


log 


e 


Th 


= -1.35 


log 


e 


Th 


= -1.15 




4018.5 



4019.0 



4019.5 
Wavelength(A) 



4020.0 



Fig. 16. — Spectral synthesis of the Th line at 4019 A in CS 31078-018. See the electronic edition of the Journal for a color version of this figure. 



40 




4500 5000 5500 6000 6500 




4500 5000 5500 6000 6500 
T,„(K) 




4500 5000 5500 6000 6500 
T,,(K) 




4500 5000 5500 6000 6500 
T„(K) 



Fig. 17. — Values of loge(Li) and [C, N, 0/Fe] vs. Teff. The symbols are as in figure|5] See the electronic edition of the Journal for a color version of this 
figure. 



41 



z 

+ 



+ 




4500 



5000 



5500 6000 
T.„ (K) 



6500 



Fig. 18. — [(C+N)/Fe] vs. [Fe/H] and Tjff. The large number of upper limits obscures any potential trend with [Fe/H]; however, it seems clear that [(C+N)/Fe] 
is not correlated with Tetf. The rms scatter for the measured [C+N/Fe] is 0.27 dex, and the average value is 0.39 dex. 




Fig. 19. — Value s of loge(Li) vs. luminosity. The symbols are as in figure|5] with the addition of the squares (colored purple in the electronic edition) from 
IGratton et alj <2000h . See the electronic edition of the Journal for a color version of this figure. 



42 




Fig. 20. — [Ti/Fe] and [Ti II/Fe] versus Teff. We also plot [Ti II/T I], which shows t hat, al though there is an offset between these values, the trend exists for 
both the neutral and singly ionized states. We also plot data taken from lPreston et al.1 )200fl) as the squares (colored purple in the electronic edition). See the 
electronic edition of the Journal for a color version of this figure. 



43 




4500 5000 5500 6000 6500 
T.„(K) 




4500 5000 5500 6000 6500 
T„(K) 




4500 5000 5500 6000 6500 
T„(K) 



Fig. 21 . — [Cr/Fe] and [Cr II/Fe] vs. Teff. We also plot [Cr Il/C r I], which shows that although there is an offset between these values, the trend exists for both 
the neutral and singly ionized states. The symbols are as in figure l20l 



-3 



-5 





1 i i 1 i i 


1 


'1 1 ' 




/ ■ 




























A 


• 


/ 

/ 












/ 










□ 


/ 










/ 












/ Cohen et al. 2008 










/ 


Francois et al. 2007 










/ 


Honda et al. 2004 


A 








/ 




□ 










Aoki el al. 2005 






• ^ 


)K 


Fulbright et al. 2004 


▼ 










This Study 


• 






















■ 1 1 ■ 






6 


-5 




-4 -3 







[Ba/H] 



Fig. 22. — [Sr/H] vs. [Ba/H]. I n these neutron-capture deficient objects, it seems that there are at least two sites that produce [Sr/H]. The main r-process line, 
as given in lSimmerer et alj 120041) is plotted as the dashed line. See the electronic edition of the Journal for a color version of this figure. 




Fig. 23. — Measured neutron-capture elements in the starCS 31078-01! 
120041) . and have been scaled to match our Eu abundance. 



The solar system r-process lines come from lArlandini et ai]n999l) and lSimmerer et sd\ 



0.0 



I ' ' ' I ' ' ' I ' ' ' I ' ' ' I ' ' ' I 




-0.8 1 ■ I ■ ■ ■ I ■ ■ ■ I ■ ■ ■ I ■ ■ ■ I ■ ■ ■ I ■ ■ ■ 

-3.2 -3.0 -2.8 -2.6 -2.4 -2.2 -2.0 
[Fe/H] 



Fig. 24. — Measured log e(Th/Eu) abundances of metal-poor stars. The symbols represent the following: the filled star is CS 31078-018 from this study, the 
diamonds are stars from Honda et al. ( 2004'), the plus sig ns are stars from Johnson & Bolte (2001), the solid circle is CS 31082-001 fromjHill et al. (2002), the 
upward-pointing triangle is HD 221170 from Ivans et al. (2006), the downward-pointing triangle is HE 1523-0901 from Frebel et al. ( 2007a), the square is CS 
22892-052 from Sneden et al. (2003), the cross is BD+17 3248 from Cowan et al. (2002), the open circle is CS 29497-004 from Christlieb et al. (2004), and the 
asterisk is HD 1 15444 from iWestin et aU (2000). The solid lines connect points that are repeated measurements of the same object. iHonda et aU I2004D suggest 
that the discrepancy in the measurements for HD 1 15444 arise from a combination of differing atm ospheric parameters and linelists. There is a clear distribution 
of values, although a majority of the stars have log e(Th/Eu)~ -0.6. The production ratio from lKratz et alil2007l) is plotted as the dashed line. 



45 




30 40 
Atomic number 



Fig. 25. — Abundance spread of our sample. When both the ionized and neutral species of an element are measured, we plot only the results for the ionized 
species. The points plotted for Fe are the Fe 11 - Fe I values of our stars, (a) Here we plot all of our measitfements for the sample from C through Eu. (b) Instead 
of all of the measurements, we show the average measured abimdance for each element. The error bars represent the rms of the abundances of each respective 
element. 



46 



(a) 




5 10 15 20 25 30 

Atomic number 



(b) 




5 10 15 20 25 30 

Atomic number 



(c) 



1.0 




„ CN ^ 




0.5 
0.0 
-0.5 




V V * V V Ni T 

F ♦ Al t 


. 


-1.0 
-1.5 




Na Mn 

Salpeter IMF, r= 1.35, x'=0 e64 
E = 0.6, E"''''= -0.5, mixing = 0.0251 i 


1 c 



5 10 15 20 25 30 

Atomic number 



Fig. 26. — Average abundance pattern of our sample fitted to the'Heger & Wooslev l'200f?) models assuming a Salpeter IMF. The filled circles are the averaged 
abundances, with the error bars corresponding to the rms of the abundance ratios over our sample. The smaller filled circles of Si and Ti represent the smaller 
weights attributed to them in the fitting procedure. The open circles are when that particular abundance is not used in the fit, and the open triangles at Sc and Zn 
represent treating the model yields as lower limits. 



(a) 




5 10 15 20 25 30 

Atomic number 



(b) 




5 10 15 20 25 30 

Atomic number 



Fig. 27. — Average abundance pattern of our sainple fitted to the IHeger & WooslevI 120081) models. The syinbols are as in Fig. 26. {a) Best fit assuming a 
Gaussian IMF. (h) Best fit to a single SN. 



47 




Fig. 28. — Abundance pattern of CS 30336-049 fitted to the'Heg er & WooslevI 120081) models. Oxygen and copper are ignored, and the Cr has been increased 
by 0.3 dex. The type of fit and its parameters are listed on each plot. 



48 



(a) 



OO OO o o 



= 0.53 O 
= 1.42 ' 



10 20 30 40 50 60 

Mass (Mq) 

(b) 



10 



o 



O (M) O OO 


®o O - 


O (0(31)01® o 












OQDOl® " 




t^QQao o 




■ tt)o 


/ = 0.42 O 






= 3.09 " 









10 20 30 40 50 60 

Mass (Mq) 




Fig. 29. — The 1000 best-fit single explosion models to (a) our average abundance pattern and (b) CS 30336-049. The values are represented linearly by 
the size of the circles, with the minimum and maximum of the fits shown in the legends of each respective plot. Mixing values are not differentiated in these 
plots. The best 50 fits for each case are plotted in red. 



(a) 




5 10 15 20 25 30 

Atomic number 




5 10 15 20 25 30 

Atomic number 



Fig. 30. — Plot of the effect of positioning the piston at the Ye boundary (edge of the iron core) and the S4 (S/NAk = 4.0) boundary (base of the convective 
shell). The black points with error bars again represent the average abundance pattern of our sample, (a) Best single star fit assuming a piston location at the edge 
of the iron core, (b) Model with the same parameters, but with the piston location at the base of the convective shell.
Internet Archive Audio

Featured

Top

Images

Featured

Top

Software

Featured

Top

Books

Featured

Top

Video

Featured

Top

Mobile Apps

Browser Extensions

Archive-It Subscription

Save Page Now

Full text of "Detailed Abundances for 28 Metal-poor Stars: Stellar Relics in the Milky Way"

See other formats
